supervised_learning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>.container { width:80% !important; }</style>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from IPython.core.display import display, HTML\n",
    "display(HTML(\"<style>.container { width:80% !important; }</style>\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def display_progress_bar(n_tot, n_prog) :\n",
    "    from time import sleep\n",
    "    progress = math.ceil((n_prog + 1) * 100 / n_tot)\n",
    "    sys.stdout.write('\\r')\n",
    "    sys.stdout.write(\"[%-100s] %d%%\" %('='*progress, progress))\n",
    "    sys.stdout.flush()\n",
    "    sleep(0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def func(date_time) :\n",
    "    \"\"\"\n",
    "    Routine converting time of day to number of seconds\n",
    "    \"\"\"\n",
    "    seconds = date_time.hour * 3600 + date_time.minute * 60 + date_time.second\n",
    "    return seconds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def display_metrics(description, model, X_test, y_test) :\n",
    "    \"\"\"\n",
    "    Routine to automate metrics calculation and display\n",
    "    \"\"\"\n",
    "    prediction = model.predict(X_test)\n",
    "    # Get Root Mean Squared Error\n",
    "    mse = metrics.mean_squared_error(y_test, prediction)\n",
    "    rmse = math.sqrt(mse)\n",
    "    # Get Mean Absolute Error\n",
    "    mae = metrics.mean_absolute_error(y_test, prediction)\n",
    "    # Get Dtermination Coefficient\n",
    "    r2 = metrics.r2_score(y_test, prediction)\n",
    "    # Display\n",
    "    print(description + \" : RMSE : %.2f, MAE : %.2f, R² : %.2f\" %(rmse, mae, r2))\n",
    "    return prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def pred_mse(test, predicted) :\n",
    "    \"\"\"\n",
    "    Looking at each prediction and computing mse/rmse\n",
    "    Return statistics on mse prediction\n",
    "    \"\"\"\n",
    "    predict_df = pd.DataFrame()\n",
    "    predict_df.loc[:, 'target'] = test\n",
    "    predict_df.loc[:, 'pred'] = predicted\n",
    "    predict_df.loc[:, 'mse'] = predict_df.loc[:, 'pred'].sub(predict_df.loc[:, 'target'], axis=0).apply(lambda x : math.pow(x, 2))\n",
    "    return predict_df, predict_df.loc[:, 'mse'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def grid_search(X_train, y_train, X_test, y_test, model, n_alphas, logmin, logmax) :\n",
    "    \"\"\"\n",
    "    Grid search function\n",
    "    \"\"\"\n",
    "\n",
    "    # Set n alphas for a grid search crossed-validation\n",
    "    alphas = np.logspace(logmin, logmax, n_alphas)\n",
    "    \n",
    "    # Initialisations\n",
    "    count=0                              # For progression display\n",
    "    best_rmse = 9999                     # Best RMSE of the grid search\n",
    "    best_alpha = 0                       # Best Alpha of the grid search\n",
    "    best_ridge = copy.deepcopy(model)    # Best model of the grid search\n",
    "\n",
    "    # Grid search\n",
    "    for a in alphas :\n",
    "        # Set model\n",
    "        model.set_params(alpha=a)\n",
    "        model.fit(X_train, y_train)\n",
    "        # Get RMSE\n",
    "        mse_temp = metrics.mean_squared_error(y_test, model.predict(X_test))\n",
    "        rmse_temp = math.sqrt(mse_temp)\n",
    "        # Look for best alpha parameter\n",
    "        if rmse_temp < best_rmse :\n",
    "            best_alpha = a\n",
    "            best_ridge = copy.deepcopy(model)\n",
    "\n",
    "        display_progress_bar(len(alphas),count)\n",
    "        count+=1\n",
    "    \n",
    "    return best_ridge, best_alpha"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def time_cross_validation(model, X, y, n_alphas, n_splits, gs=True) :\n",
    "    \"\"\"\n",
    "    Cross validation function with a grid search over alpha parameter for Ridge Regression\n",
    "    \"\"\"\n",
    "    \n",
    "    # Create time splits\n",
    "    tscv = model_selection.TimeSeriesSplit(n_splits=n_splits)\n",
    "\n",
    "    # Initialise outputs\n",
    "    errors_cv = []\n",
    "    alphas_cv = []\n",
    "    \n",
    "    # Initialise splits count\n",
    "    num_split = 1\n",
    "\n",
    "    # Time cross--validation\n",
    "    for train_index, test_index in tscv.split(X):\n",
    "        # Create train and test sets\n",
    "        X_train_cv, X_test_cv = X.iloc[train_index, :], X.iloc[test_index, :]\n",
    "        y_train_cv, y_test_cv = y.iloc[train_index], y.iloc[test_index]\n",
    "\n",
    "        if gs :\n",
    "            best_model, best_alpha = grid_search(X_train_cv, y_train_cv, X_test_cv, y_test_cv, model,\n",
    "                                                 n_alphas, 1, 10)\n",
    "        else :\n",
    "            model.fit(X_train_cv, y_train_cv)\n",
    "            best_model = model\n",
    "            best_alpha = NaN\n",
    "        \n",
    "        # Get split model errors with best grid search model\n",
    "        mse_cv = metrics.mean_squared_error(y_test_cv, best_model.predict(X_test_cv))\n",
    "        rmse_cv = math.sqrt(mse_cv)\n",
    "        mae_cv = metrics.mean_absolute_error(y_test_cv, best_model.predict(X_test_cv))\n",
    "        r2_cv = metrics.r2_score(y_test_cv, best_model.predict(X_test_cv))\n",
    "        # Save errors\n",
    "        errors_cv.append([mae_cv, rmse_cv, r2_cv])\n",
    "        # Save best alpha for this time split\n",
    "        alphas_cv.append(best_alpha)\n",
    "        # Display\n",
    "        display_metrics(\"\\nModel regression for split number \" + str(num_split), best_model, X_test_cv, y_test_cv)\n",
    "        # Loop\n",
    "        num_split += 1\n",
    "        \n",
    "    return errors_cv, alphas_cv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1) Importations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline\n",
    "import pandas as pd\n",
    "import os\n",
    "import sys\n",
    "import seaborn as sns\n",
    "from sklearn import preprocessing, metrics, linear_model, model_selection, dummy\n",
    "import scipy as sp\n",
    "import math\n",
    "import time\n",
    "import copy\n",
    "from dateutil import parser\n",
    "import pickle"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2) Data loading"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Getting current path\n",
    "path = os.getcwd()\n",
    "\n",
    "# Verifying data presence\n",
    "try :\n",
    "#     data = pd.read_csv(path + \"/data/data_airlines.csv\")\n",
    "#     with open(path + '/data/data_airlines.pkl', 'rb') as f :\n",
    "    with open(path + '/data/data_more_features.pkl', 'rb') as f :\n",
    "        data = pickle.load(f)\n",
    "except FileNotFoundError :\n",
    "    print(\"Please check if the file 'data_airlines.csv' is in the 'data' folder at the current location\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3) Data preparation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3.1) Overview"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4518157, 9)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ARR_DELAY_NEW</th>\n",
       "      <th>DEST_AIRPORT_ID</th>\n",
       "      <th>DISTANCE</th>\n",
       "      <th>ORIGIN_AIRPORT_ID</th>\n",
       "      <th>TAIL_NUM</th>\n",
       "      <th>UNIQUE_CARRIER</th>\n",
       "      <th>DATE_TIME</th>\n",
       "      <th>TIME_SECONDS</th>\n",
       "      <th>ARR_TIME_SECS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5480165</th>\n",
       "      <td>10.0</td>\n",
       "      <td>11298.0</td>\n",
       "      <td>1235.0</td>\n",
       "      <td>12892</td>\n",
       "      <td>N860AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>2016-01-01 00:10:00</td>\n",
       "      <td>600</td>\n",
       "      <td>18096.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5482988</th>\n",
       "      <td>0.0</td>\n",
       "      <td>11057.0</td>\n",
       "      <td>1773.0</td>\n",
       "      <td>14107</td>\n",
       "      <td>N157UW</td>\n",
       "      <td>AA</td>\n",
       "      <td>2016-01-01 00:15:00</td>\n",
       "      <td>900</td>\n",
       "      <td>21696.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5476021</th>\n",
       "      <td>0.0</td>\n",
       "      <td>11057.0</td>\n",
       "      <td>2296.0</td>\n",
       "      <td>14771</td>\n",
       "      <td>N558UW</td>\n",
       "      <td>AA</td>\n",
       "      <td>2016-01-01 00:15:00</td>\n",
       "      <td>900</td>\n",
       "      <td>29280.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5661633</th>\n",
       "      <td>0.0</td>\n",
       "      <td>13930.0</td>\n",
       "      <td>1846.0</td>\n",
       "      <td>14771</td>\n",
       "      <td>N472UA</td>\n",
       "      <td>UA</td>\n",
       "      <td>2016-01-01 00:24:00</td>\n",
       "      <td>1440</td>\n",
       "      <td>25344.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5444953</th>\n",
       "      <td>0.0</td>\n",
       "      <td>13303.0</td>\n",
       "      <td>2342.0</td>\n",
       "      <td>12892</td>\n",
       "      <td>N864AA</td>\n",
       "      <td>AA</td>\n",
       "      <td>2016-01-01 00:30:00</td>\n",
       "      <td>1800</td>\n",
       "      <td>31008.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         ARR_DELAY_NEW  DEST_AIRPORT_ID  DISTANCE  ORIGIN_AIRPORT_ID TAIL_NUM  \\\n",
       "5480165           10.0          11298.0    1235.0              12892   N860AA   \n",
       "5482988            0.0          11057.0    1773.0              14107   N157UW   \n",
       "5476021            0.0          11057.0    2296.0              14771   N558UW   \n",
       "5661633            0.0          13930.0    1846.0              14771   N472UA   \n",
       "5444953            0.0          13303.0    2342.0              12892   N864AA   \n",
       "\n",
       "        UNIQUE_CARRIER           DATE_TIME  TIME_SECONDS  ARR_TIME_SECS  \n",
       "5480165             AA 2016-01-01 00:10:00           600        18096.0  \n",
       "5482988             AA 2016-01-01 00:15:00           900        21696.0  \n",
       "5476021             AA 2016-01-01 00:15:00           900        29280.0  \n",
       "5661633             UA 2016-01-01 00:24:00          1440        25344.0  \n",
       "5444953             AA 2016-01-01 00:30:00          1800        31008.0  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(data.shape)\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Some feature are categorical : need for encoding"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3.2) Encoding categorical features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Label encoding airlines"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "le = preprocessing.LabelEncoder()\n",
    "label_airlines = le.fit_transform(data.UNIQUE_CARRIER)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One Hot encoding airlines from label encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "oh_airlines = pd.get_dummies(label_airlines)\n",
    "# Reindex\n",
    "oh_airlines.index = data.index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Label encoding aircrafts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "le = preprocessing.LabelEncoder()\n",
    "label_aircrafts = pd.DataFrame(le.fit_transform(data.TAIL_NUM), columns=['TAIL_NUM'])\n",
    "# Reindex\n",
    "label_aircrafts.index = data.index\n",
    "# Concat\n",
    "data_reg = pd.concat((oh_airlines, label_aircrafts), axis=1)\n",
    "# Relieve Memory\n",
    "del label_airlines, oh_airlines, label_aircrafts"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One Hot encoding departure airports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "oh_dep_airport = pd.get_dummies(data.ORIGIN_AIRPORT_ID)\n",
    "# Concat\n",
    "data_reg = pd.concat((data_reg, oh_dep_airport), axis=1)\n",
    "# Relieve Memory\n",
    "del oh_dep_airport"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One Hot encoding arrival airports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# oh_arr_airport = pd.get_dummies(data.DEST_AIRPORT_ID)\n",
    "# # Concat\n",
    "# data_reg = pd.concat((data_reg, oh_arr_airport), axis=1)\n",
    "# # Relieve Memory\n",
    "# del oh_arr_airport"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Standardise numerical features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "std_scale = preprocessing.StandardScaler().fit(data.loc[:, ['TIME_SECONDS', 'ARR_TIME_SECS', 'DISTANCE']].values)\n",
    "data_num = pd.DataFrame(std_scale.transform(data.loc[:, ['TIME_SECONDS', 'ARR_TIME_SECS', 'DISTANCE']].values), columns=['DEP_TIME_SECS_STD', 'ARR_TIME_SECS_STD', 'DISTANCE_STD'])\n",
    "# Reindex\n",
    "data_num.index = data.index\n",
    "# Concat\n",
    "data_reg = pd.concat((data_reg, data_num), axis=1)\n",
    "# Relieve Memory\n",
    "del data_num"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Concatenate everything into processed data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>15380</th>\n",
       "      <th>15389</th>\n",
       "      <th>15412</th>\n",
       "      <th>15624</th>\n",
       "      <th>15919</th>\n",
       "      <th>16218</th>\n",
       "      <th>DEP_TIME_SECS_STD</th>\n",
       "      <th>ARR_TIME_SECS_STD</th>\n",
       "      <th>DISTANCE_STD</th>\n",
       "      <th>ARR_DELAY_NEW</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5480165</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-2.703141</td>\n",
       "      <td>-1.992056</td>\n",
       "      <td>0.538910</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5482988</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-2.686161</td>\n",
       "      <td>-1.799513</td>\n",
       "      <td>1.388456</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5476021</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-2.686161</td>\n",
       "      <td>-1.393889</td>\n",
       "      <td>2.214315</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5661633</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-2.655597</td>\n",
       "      <td>-1.604403</td>\n",
       "      <td>1.503729</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5444953</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-2.635220</td>\n",
       "      <td>-1.301469</td>\n",
       "      <td>2.286953</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 218 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         0  1  2  3  4  5  6  7  8  9      ...        15380  15389  15412  \\\n",
       "5480165  1  0  0  0  0  0  0  0  0  0      ...            0      0      0   \n",
       "5482988  1  0  0  0  0  0  0  0  0  0      ...            0      0      0   \n",
       "5476021  1  0  0  0  0  0  0  0  0  0      ...            0      0      0   \n",
       "5661633  0  0  0  0  0  0  0  0  0  1      ...            0      0      0   \n",
       "5444953  1  0  0  0  0  0  0  0  0  0      ...            0      0      0   \n",
       "\n",
       "         15624  15919  16218  DEP_TIME_SECS_STD  ARR_TIME_SECS_STD  \\\n",
       "5480165      0      0      0          -2.703141          -1.992056   \n",
       "5482988      0      0      0          -2.686161          -1.799513   \n",
       "5476021      0      0      0          -2.686161          -1.393889   \n",
       "5661633      0      0      0          -2.655597          -1.604403   \n",
       "5444953      0      0      0          -2.635220          -1.301469   \n",
       "\n",
       "         DISTANCE_STD  ARR_DELAY_NEW  \n",
       "5480165      0.538910           10.0  \n",
       "5482988      1.388456            0.0  \n",
       "5476021      2.214315            0.0  \n",
       "5661633      1.503729            0.0  \n",
       "5444953      2.286953            0.0  \n",
       "\n",
       "[5 rows x 218 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data_reg = pd.concat((data_reg, data.loc[:, ['ARR_DELAY_NEW']]), axis=1)\n",
    "# Relieve Memory\n",
    "del data\n",
    "display(data_reg.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Set Target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Target : arrival delay\n",
    "y = data_reg.loc[:, 'ARR_DELAY_NEW']\n",
    "X = data_reg.loc[:, data_reg.columns  != 'ARR_DELAY_NEW']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Split in train/test sets : Learn over 70% of the year to predict the future 30% of the year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_size = 0.3\n",
    "n_train = math.floor(X.shape[0] * (1 - test_size))\n",
    "\n",
    "X_train = X.iloc[:n_train, :]\n",
    "X_test = X.iloc[n_train:, :]\n",
    "\n",
    "y_train = y.iloc[:n_train]\n",
    "y_test = y.iloc[n_train:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Save Memory"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "del data_reg#,X, y, "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4) First models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4.1) Dummy model : Predicting the mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "dum = dummy.DummyRegressor(strategy='mean')\n",
    "\n",
    "# Entraînement\n",
    "dum.fit(X_train, y_train)\n",
    "\n",
    "# Prédiction sur le jeu de test\n",
    "y_pred_dum = dum.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dummy mean prediction on training set : RMSE : 21.13, MAE : 12.87, R² : 0.00\n",
      "Dummy mean prediction on testing set : RMSE : 19.87, MAE : 12.28, R² : -0.00\n"
     ]
    }
   ],
   "source": [
    "# Get the errors\n",
    "display_metrics(\"Dummy mean prediction on training set\", dum, X_train, y_train)\n",
    "display_metrics(\"Dummy mean prediction on testing set\", dum, X_test, y_test)\n",
    "# Get the mse composition\n",
    "pred_df_dum, mse_stats_df_dum = pred_mse(y_test, y_pred_dum)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Getting mean MSE for 2-min delay windows\n",
    "delays = []\n",
    "mean_mse_delay_dum = []\n",
    "delay = 0\n",
    "for i in range(70) :\n",
    "    mean_mse_delay_dum.append(pred_df_dum.loc[:, 'mse'][pred_df_dum.loc[:, 'target'].between(delay, delay+2)].mean())\n",
    "    delays.append(delay)\n",
    "    delay += 2\n",
    "sorted_pred_df_dum = pred_df_dum.sort_values(by=['mse'], ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAg8AAAFFCAYAAABrDg41AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xl8VNX9//HXh32RTQVEloIVW9dazFdQq6IWpSogLogbqChFpO4LKluhKK0LrdWfra0WrAsiasWlWkUURUHABQRkDyEJkEDIQvZkzu+Pe4OBmYQEkjuTzPv5eMxjZs45987nOjH5cM6555hzDhEREZGqahDtAERERKRuUfIgIiIi1aLkQURERKpFyYOIiIhUi5IHERERqRYlDyIiIlItSh5ERESkWpQ8iIiISLUoeRAREZFqaRTtAGLB4Ycf7rp37x7tMERERAKxbNmyHc659gd6vJIHoHv37ixdujTaYYiIiATCzDYfzPEathAREZFqUfIgIiIi1RJo8mBmz5tZmpl9X67sVTP71n8kmtm3fnl3M8svV/e3csecYmYrzGy9mT1pZuaXH2pmH5rZOv+5XZDXJyIiEg+C7nmYAfQvX+Ccu9I5d7Jz7mTgdeCNctUbyuqcc6PKlT8DjAR6+o+yc44F5jnnegLz/PciIiJSgwJNHpxzC4CMSHV+78EQ4JXKzmFmnYDWzrkvnXMOeAG4xK8eBMz0X88sVy4iIiI1JJbmPJwJbHfOrStX1sPMvjGzT83sTL+sM5Bcrk2yXwbQ0Tm3FcB/7lDbQYuIiMSbWLpV8yr27nXYCnRzzu00s1OA/5jZ8YBFONZV98PMbCTe0AfdunWrtG12djZpaWkUFxdX92OknmjcuDEdOnSgdevW0Q5FRCTqYiJ5MLNGwKXAKWVlzrlCoNB/vczMNgDH4PU0dCl3eBcg1X+93cw6Oee2+sMbaRV9pnPuWeBZgISEhAqTj+zsbLZv307nzp1p3rw5/txMiSPOOfLz80lJSQFQAiEicS9Whi1+DfzgnNszHGFm7c2sof/6KLyJkRv94YgcM+vjz5MYBrzlHzYXGO6/Hl6u/IClpaXRuXNnWrRoocQhTpkZLVq0oHPnzqSlVZiPiojEjaBv1XwF+BL4mZklm9kIv2oo4RMlzwKWm9l3wBxglHOubLLlLcA/gfXABuC/fvk0oJ+ZrQP6+e8PSnFxMc2bNz/Y00g90Lx5cw1diUid133suzQ54uhT9t+yYoEOWzjnrqqg/PoIZa/j3boZqf1S4IQI5TuB8w4uynDqcRDQz4GI1DG7E+GQ7rVy6lgZthAREZGakrEM3j0elt0BodIaP72ShzgwadIkzIyePXtGrD/66KMxMyZNmhRsYPtx/fXXY2b069cvrC4/P59WrVphZsyYMWNPeXFxMU888QQnnHACLVq04PDDD6d3795Mm/bjCFZiYiJmFvHx61//OohLExGpPblJ8MnFUJoHa/4Cnw2G4t17qhOnXUTRtvXLDuYjYuJuC6l9zZo1Y9OmTSxdupSEhIQ95UuWLGHz5s00a9YsitFV7JBDDmH+/Pls376djh077il/5513IrYfM2YML730Eg899BC9e/cmMzOTRYsW8fbbbzN27N4Ljj722GOcccYZe5W1adOm5i9CRCQoxdnw6cVQsO3HspS3+f6Fk7lh0yTSS9qROO2ig/4YJQ9xomXLlvTq1YtZs2btlTzMmjWLc889l2XLDioJrTU/+9nPyMnJ4bXXXmPMmDF7ymfNmsXAgQN5+eWX95Tl5eXxr3/9i6lTp3LvvffuKb/00kvxFiMNP3efPn1q9wJERIISKoHPr4TMFWFVDSklL9S0xj5KwxZxZOjQocyePXvPH1LnHLNnz2bo0KER23/++eecffbZtGjRgsMOO4ybb76ZnJycPfVbt27lxhtv5KijjqJ58+Ycc8wxjBs3jqKioj1tyoYIZs+ezW9/+1vatGlDly5dmDhxIqFQqEpxX3nllcyaNWvP+5ycHN57772wuHNzcykuLuaII44IO4cmO4pIveYcLP0dbH0/vK55J25InERuqEWNfZx6Hqrr5Rj6I3R19RbWvPTSS7nlllv4/PPPOfPMM/nss89IT09n8ODBe/1LHWDhwoWcd955XHLJJcyZM4edO3cyduxYdu3axZw5cwDYsWMHhx56KE888QTt2rVj7dq1TJo0ifT0dP7+97/vdb777ruPyy67jDlz5jBv3jwmT57M8ccfz5AhQ/Yb91VXXcUf/vAHkpKS6NatG2+++Sbt2rXj7LPP3qtd+/bt6dq1K5MmTaJly5b069ePVq1aVXjeUChESUnJXmUNGjSgQQPl1CJSx/wwHdb/Lby8YQs4+x0WDe5Vox+n35JxpG3btvTv33/Pv+JnzZpF//79adu2bVjbsWPHcvrpp/Pqq69y4YUXct111/Hvf/+b119/ne+/93ZUP/HEE3nssce45JJLOPvss7nhhht49NFHmTlz5l69DwBnnXUWjz/+OP369WPatGn84he/4I033gj73EiOPfZYTjzxRF599dU9cQ8ZMiTiH/kZM2aQk5PDZZddRtu2bUlISOCxxx4Liwdg0KBBNG7ceK/H5MmTqxSTiEjM2PIf+OaesOKQM0asv5vuf9pa4x+p5CHODB06lDlz5lBYWMicOXMiDlnk5eXx5ZdfMmTIEEpKSvY8fvWrX9G4ceM98yOcc/z5z3/muOOOo3nz5jRu3JhrrrmGwsJCkpKS9jrn+eefv9f74447juTkZKpq6NChzJo1i4yMDD766KMKh1rOPfdcNmzYwCuvvMKNN97Izp07uffeezn33HPDhkmmT5/OkiVL9nqMHDmyyjGJiETdzqXwxdVE2uJpcurNzMvpXSsfq+QhzgwcOJDdu3fz0EMPkZuby4ABA8La7Nq1i9LSUkaPHr3Xv8qbNm1KcXExW7ZsAeDPf/4zd999N4MHD+att97iq6++4umnnwagoKBgr3Pu27vRpEmTsDaVGTp0KF9//TUPP/wwnTt3rnSiY6tWrRg6dCj/+Mc/2LhxI+PHj2fhwoW8/fbbe7U7+uijSUhI2Otx5JFHVjkmEZGoyk2CTwdAaX543TFjmLFzYK19tOY8xJmWLVty8cUXM336dK644gpatmwZ1qZt27Z71n248MILw+rL/sC+9tprXHHFFUydOnVP3apVq2ol7h49enDqqacyffr0sPkZlTEz7r33XqZMmcIPP/zAoEGDaiU+EZFAFWXCJxfufUtmmSMvgl7TSUyovT/xSh6qq5qTFGPRLbfcQmFhIaNGjYpY37JlS/r06cOaNWuYMGFChefJz8+nadO9b/156aWXajTW8u6++25efvllhg0bFrG+uLiY3NzcsF6OdevWAey1ToSISJ1VWgSfXQZZK8OqVuYfxRX/u5689z+okfUcKqLkIQ717duXvn37VtrmT3/6E+eddx4NGjTg8ssvp1WrViQlJfHuu+8ydepUjjnmGPr168eTTz5J7969+elPf8pLL73E+vXray3uIUOGVHp3RlZWFscccwzDhw/nnHPOoU2bNqxZs4ZHHnmEzp07M3jw4L3ar1mzhsMPP3yvsmbNmnHyySfXSvwiIgfNOfjqZtj+cXhd8yMZsXoCeaHa38xRyYNE9Ktf/YoFCxYwceJErrvuOkpLS/nJT35C//799/wLfsKECaSnpzNu3DjAuxX0ySefjDiPIgitW7fmvvvu47333uPll18mOzubzp07c8EFFzBu3Liw1SPvuSd8dvJPf/rTWk2AREQOyorfw6YXwssbHQJnv8O2xamBhGGRVt6LNwkJCW7p0qUR61avXs2xxx4bcEQSq/TzICJRs3EGLLohrLjENWDEponMHFfxMPO+zGyZcy5h/y0j090WIiIisW7bR7D45ohV41JG8+nuUwINR8mDiIhILMtc4U2QdCVhVU9tH8KsjP6Bh6Q5DyIiIrEqd4t3S2ZxdnjdT65mzFUvMiYKe/eo50FERCQWFWbA/AsgL3w13kW7T+CYd66AKG36p+RBREQk1pTkeatHZq8Oq1pX0JWRm8dR5BpHITCPkgcREZFYEiqBz6+EHV+E1zXryA2bJpFdekjwcZWjOQ8iIiKxwjn46reQ+k5YVU5pc65c/hCfT7kxCoHtTT0PIiIisWL5ONj4fFhxYagRIxPHs6rgqCgEFU7Jg4iISCxY8ySsfDhChXHHlnv4MvekwEOqiIYtREREoi3pNVh2R+S6hKd45urRwcazH+p5iBOTJk3CzOjZs2fE+qOPPnrPNtxlZsyYwSmnnEKrVq1o164dv/zlL7nrrrv2Os7MIj6OPvro2rwcEZH6I20BfHEtEL5dxJPbr6T78z8JPqb9UM9DHGnWrBmbNm1i6dKlJCT8uKT5kiVL2Lx5M82aNdtT9sgjjzB+/Hjuu+8+pk2bRkFBAcuWLePFF1/kiSee2Ou8d999N5dffnnYZ4mIyH5krYJPB0GoKKzqlZ3n88T2a6MQ1P4peYgjLVu2pFevXsyaNWuv5GHWrFmce+65LFu2bE/ZU089xW9/+1sefvjH8bcBAwYwceLEsPN2796dPn361G7wIiL1TV4qzP8NFGeG13UeyLjlI4DoLAK1P4EOW5jZ82aWZmbflyubZGYpZvat/7iwXN0DZrbezNaY2QXlyvv7ZevNbGy58h5mttjM1pnZq2bWJLirqxuGDh3K7NmzKdtN1TnH7NmzGTp06F7tMjMzOeKII8KOtyitZiYiUq8U58CnF0FeUnjdYb3hjFfYMG0gidMuInHaRcHHtx9Bz3mYAUTawWO6c+5k//EegJkdBwwFjveP+X9m1tDMGgJPA78BjgOu8tsC/NE/V09gFzCiVq+mDrr00kvZvn07n3/+OQCfffYZ6enpDB48eK92vXr14q9//SszZ85k586dlZ4zFApRUlKy1yMUCtXaNYiI1GmhYvjsctj1bVjVpsJO9FpwG93HzY9CYFUXaPLgnFsAZFSx+SBglnOu0Dm3CVgPnOo/1jvnNjrnioBZwCDz/kl8LjDHP34mcEmNXkA90LZtW/r378+sWbMAb8iif//+tG3bdq92Tz/9NIcccgjXX3897du35/jjj2fChAlkZ4dvznL77bfTuHHjvR433hj9RUxERGJO2SJQ2/4XXte0PcM3TSajtE3wcVVTrNxtMcbMlvvDGu38ss7AlnJtkv2yisoPAzKd27NnaVl5THl58Wb6PDKPlxdvjloMQ4cOZc6cORQWFjJnzpywIQuAk046idWrVzN37lxGjx6Nc44pU6aQkJDA7t2792p77733smTJkr0e5e/aEBER34pJsPFf4eUNm8PZ75BU1CnwkA5ELEyYfAaYgnePyhTgceBGIs8ScUROeFwl7SMys5HASIBu3bpVL+KD8OTH69mWVcBfP17P1b2jc/vNwIEDuemmm3jooYfIzc1lwIABEds1bdqUAQMG7Kl/7rnnuOmmm3juuee4/fbb97Tr1q3bXhMwRUQkgjVPwfeTw4pLXQNGrr+Hed+kx+T8hkii3vPgnNvunCt1zoWAf+ANS4DXc9C1XNMuQGol5TuAtmbWaJ/yij73WedcgnMuoX379jVzMVVw27lH06lNM353bvTWQWjZsiUXX3wx06dPZ8CAAbRs2bJKx40YMYJDDz2UH374oZYjFBGpZxJfgWW3Rawan3IL83J6BxzQwYl6z4OZdXLObfXfDgbK7sSYC7xsZk8ARwI9ga/wehh6mlkPIAVvUuXVzjlnZvOBy/HmQQwH3gruSqrm6t4/iVqPQ3m33HILhYWFjBo1KmJ9WloaHTp02KssPT2drKwsOnbsGESIIiL1Q+oH8OUwInaGH/cALy8/I/CQDlagyYOZvQL0BQ43s2RgItDXzE7G+6+aCPwWwDm30sxmA6uAEuBW51ypf54xwAdAQ+B559xK/yPuB2aZ2R+Ab4DnArq0Oqdv37707du3wvoTTzyRQYMGcf7559OhQwc2b97MY489RosWLRg+fPhebRMTE1m0aNFeZWZG7951K5MWEalxOxbBZ5fCnul45Rx1I/xiKokn171b4ANNHpxzV0UorvAPvHNuKjA1Qvl7wHsRyjfy47CHHIQJEybw1ltvcdttt5GRkcERRxzB6aefzquvvkqPHj32avv444/z+OOP71XWsGFDSkoi/M8iIhIvslbBJxdBaV5Y1QdZfRj9n0Fs6FP3EgcAK1ssKJ4lJCS4pUuXRqxbvXo1xx57bMARSazSz4OIVEnuZvjfGZCfEla1aPcJDN80mULXJGoTJM1smXPugGe6R33CpIiISL1SkAbzL4iYONDuZG5OHE+hq9sLIEd9wqSIiEi9UbQLPj4fsteEVSUWduLyhXez4uEhUQisZqnnQUREpCYU58D8CyHzu7CqtOJ2XLdpCjtK2kU4sO5R8iAiInKwSgtgwSDYuSi8rnEbhm2azJai8M0G6yoNW1SBc067SQqaXCwiEYWK4bMrYHuEzawatYS+/+X9K04LPq5apJ6H/WjcuDH5+fnRDkNiQH5+Po0bN452GCISS0Kl8MV1kPpOWFVhqDFXrXmQ7o9XdT/IukPJw3506NCBlJQU8vLy9C/POOWcIy8vj5SUlLBVN0UkjrkQfDUSkl4Nr7NGjN48li9zfxF8XAHQsMV+tG7dGoDU1FSKi4ujHI1ES+PGjenYseOenwcRiXPOwdf3wMbnI1QanPZv5n3XKvCwgqLkoQpat26tPxoiIvKjlQ/DmukRq+7b8jtm/61Vndkh80Bo2EJERKQ61v0Nlo+LWDU59WZm7zo/4ICCp54HERGRqto8G5aMjlx34u95fvkpwcYTJUoeREREqiL1A/jyWiJtrf18+kAmv9SrXg9VlKdhCxERkf0p21o7FD5x/vVd5zBl601A/KwHpORBRESkMpkr4ZMLI26tTecB3L/ldlyc/TnVsIWIiEhFstfC/H7ehlf7an8mnPEq689uHnxcURZfqZKIiEhV5ayHeedA/tawqpX5R3Hi/DHQKP4SB1DyICIiEm73Jph3LuSnhlVtKuzE8E2/JyfUMgqBxQYlDyIiIuXlbvZ6HPK2hNe16MJ1G+vP1toHSnMeREREyuRugY/O8RKIfWwrPpShX4/n8ykjohBYbFHPg4iICEBeijdUkbsprCqtuB1Xb3yYxKLOUQgs9ih5EBERyd/mJQ6714fXNevA1RunsrGwS/BxxSgNW4iISHwrSIePz4OctWFVO0tac9V3E/jo97dEIbDYpZ4HERGJX4UZ8HE/yFoVVrWrpBXXbvwDawu7Bx9XjFPyICIi8akoC+ZfAJnfhdc1bsu1G//A6oKjgo+rDtCwhYiIxJ/iHPjkN5CxNLyucWs493+8e8X/BR9XHaGeBxERiS8lefDpxbDjy7Cq3NJmXLp6PN0fTYtCYHVHoMmDmT1vZmlm9n25skfN7AczW25mb5pZW7+8u5nlm9m3/uNv5Y45xcxWmNl6M3vSzMwvP9TMPjSzdf5zfK/iISIieystgAWDIG1BeF3D5tyYOJGv844NPq46JuiehxlA/33KPgROcM6dBKwFHihXt8E5d7L/GFWu/BlgJNDTf5SdcywwzznXE5jnvxcREYHSQlhwKWz7KLyuQVM46y0W554YfFx1UKBzHpxzC8ys+z5l/yv3dhFweWXnMLNOQGvn3Jf++xeAS4D/AoOAvn7TmcAnwP0HH7mIiNRppYXw2WWw9b9hVUWhRvx201jmf1tE4rSLohBc3RNrcx5uxEsCyvQws2/M7FMzO9Mv6wwkl2uT7JcBdHTObQXwnzvUdsAiIhLjSovg8ysg9d3wOmvI75LuY36OJkdWR8zcbWFmDwElwEt+0Vagm3Nup5mdAvzHzI4HLMLh7gA+byTe0AfdunU7sKBFRCS2hYph4ZWQ8nZ4nTWA017kg+9aBR9XHRcTyYOZDQcuBs5zzjkA51whUOi/XmZmG4Bj8Hoayq8R2gUo2zN1u5l1cs5t9Yc3Kpwu65x7FngWICEhodrJh4iIxLhQMSwcCsn/CasqdQ24K+lO/nLVUBKnRSG2Oi7qwxZm1h9vXsJA51xeufL2ZtbQf30U3sTIjf5wRI6Z9fHvshgGvOUfNhcY7r8eXq5cRETiSagEvrgGtrwRXuWMe7bcwVuZ50QhsPoh0J4HM3sFb0Lj4WaWDEzEu7uiKfChf8flIv/OirOAyWZWApQCo5xzGf6pbsG7c6M53hyJsnkS04DZZjYCSAKuCOCyREQkloSK4YtrIem1CJXGvcl38GbmuYGHVZ+YP0oQ1xISEtzSpRFWGRMRkbqlJL/iyZHAvVtu49H7/xJwULHHzJY55xIO9PioD1uIiIjUiOJsb8npChKH+5N/x2u7zg84qPopJiZMioiIHJSCHRXvVQE8mHwrr2ZcEHBQ9ZeSBxERqdvyUrxttbNXh1WFnPFAyhj+eN+TPByF0OorDVuIiEjdlbMBPvxVxMShKNSIMUn3q8ehFqjnQURE6qas1TDvXCjYFl7XsDk3bxzLp7tPCT6uOKDkQURE6p7stRUmDtmlLbhx/UTmTLgvCoHFByUPIiJSt+zeWGHisLOkNcM2TWZl/tFRCCx+aM6DiIjUHbmbvcQhPyW8rkUXhmz4oxKHAKjnQURE6oa8ZC9xyN0cXteiC/x6AfMu6RF8XHFIPQ8iIhL78rfCvPO8IYt9bC8+lL5fj6f7H1ZFIbD4pORBRERiW0GalzjkrA2va9aBqzdOJbGoc/BxxTElDyIiErvykuGjvhHXcaDpYXDuPDYUdg08rHinOQ8iIhKbstfAx+dDXlJYVWbJIVy9diKrlmwmcdpFUQguvqnnQUREYk/G1/DhmRETBxq35rpNU1hVcFTwcQmg5EFERGLN9k+8oYrC9PC6xq3hnA9Ykd8z6KikHA1biIhI7Eh+Cz6/EkKFYVXpxW0ZvnYy713Rh8RpUYhN9lDyICIisWHjTFg8AlxpWNWWoo5cu3EKm4uOjEJgsi8NW4iISPT9MB0WXR8xcaDN8Vy+/o9KHGKIeh5ERCR6nIPl42Hl1IjV3+T+jOtXjuO7qUMDDkwqo+RBRESiI1QKy34H656JWL0g55eM2vwgeaHmAQcm+6PkQUREgldaBF8Og6RXI9d3u4Kb3r2aItc42LikSpQ8iIhIsEry4LPLYOv7Eatf3tmfccuvZeO0gQEHJlWl5EFERIJTtAs+HQDpCyNWP512BY9uGwZYsHFJtehuCxERCUZeCnx4VoWJA798lEe3DUeJQ+xTz4OIiNS+rB9g/gURl5sudQ14IHkMs5cfq30q6gglDyIiUrt2LIZPL4LCneF1DZoweuM9fJB9evBxyQFT8iAiIrUn9X1vcmRpXnhdo0PgrDf54NvwpagltgWePJjZ88DFQJpz7gS/7FDgVaA7kAgMcc7tMjMD/gJcCOQB1zvnvvaPGQ6M80/7B+fcTL/8FGAG0Bx4D7jdOecCuTgREfnRphdh0Q3gSsKqdpS04fp1v+f7rws1VFEHRWPC5Ayg/z5lY4F5zrmewDz/PcBvgJ7+YyTwDOxJNiYCvYFTgYlm1s4/5hm/bdlx+36WiIjUttWPw5fXRUwcaNmDy9f/ie/zjw4+LqkRgScPzrkFQMY+xYOAmf7rmcAl5cpfcJ5FQFsz6wRcAHzonMtwzu0CPgT6+3WtnXNf+r0NL5Q7l4iI1DYXgmV3wTf3RKxemX8UnP8FiUWdAw5MalKszHno6JzbCuCc22pmHfzyzsCWcu2S/bLKypMjlIuISG0rLfQ2t9o8K2L1F7tPYmTiOL5vfoSGKuq4WEkeKhLpZl93AOXhJzYbiTe8Qbdu3Q40PhERASjKgs8Gw/b5EavfzTyDO7fco+Wm64lYSR62m1knv9ehE5DmlycDXcu16wKk+uV99yn/xC/vEqF9GOfcs8CzAAkJCZpQKSJyoPJS4ZPfQObyiNUzd1zE8DFvcVGDhgEHJrUlVlaYnAsM918PB94qVz7MPH2ALH944wPgfDNr50+UPB/4wK/LMbM+/p0aw8qdS0REalrWavjfaRUmDn/aOoyJqaNAiUO9Eo1bNV/B6zU43MyS8e6amAbMNrMRQBJwhd/8PbzbNNfj3ap5A4BzLsPMpgBL/HaTnXNlkzBv4cdbNf/rP0REpKalf+HtU1G07xx4wBpxd9LveH3XecHHJbXOtASCN2yxdOnSaIchIlJ3JL8FC4dCaUFYVW5pM0ZvfoCZ4yZEITCpCjNb5pxLONDjY2XOg4iI1BXr/g5LR3u3Ze5jR0kbbtg0iRX5PaMQmARFyYOIiFSNc7BiInw/JXL9IT/lsqX3s7noyGDjksApeRARkf0LlcCSUbDhuYjV3+X15MaVE1k29ZqAA5NoUPIgIiKVK8mFz4dC6jsRqz/JPoXRSWPJCzUPODCJlv3eqmlma83spHLvzcyeN7Nu+7Q71cyKaiNIERGJkvyt8FHfChMHegznpsTxShziTFV6Ho4GmpV73wBvLYan8G6rLGOAbuQVEakvdn0Hn14MeckRq5/aPoTHll9O4rSLAw5Mou1Ahy0iLQMtIiL1Rcq73q2YJbsjVBrjUkbx4k7tTxGvYmWFSRERiQXOwZonYcHAiIlDQagJnDlHiUOc04RJERHxhEpg2e2w7v9FrE4vbsvNieP5z7WXkjgt4NgkplQ1ebjMzMpWomqAt1PlFf5+E2W612RgIiISoKJdsPAq2PpBxOo1Bd0YsWkiycUdAw5MYlFVk4d7I5TdH6FMa12LiNQ1WT94wxQ56yJWf5rTi7Nv+JjPm7QJODCJVftNHpxzmhchIlJfpbwHX1wFxdkRq1/YcRG/Tx3JBiUOUo7mPIiIxCPnYPWf4NsHiNhpbA2YlHwTM3YOQDfYyb4OOHkwsxbACODnwHZgpnNuc00FJiIitaQkHxbfBJtfjlidXdqC25LuZcZDk5gUbGRSR+w3eTCzx4EBzrljypW1ApYAPYFdQBvgLjM71Tm3traCFRGRg5SXAgsugYylEas3Fh7JzYnj2VDYNeDApC6pynyGc4AX9ym7BzgGuNk5dzhwJJAIjK/R6EREpOZkLIMPTq0wcaDTBVyy/gklDrJfVRm26A4s26fsMmDZoY6EAAAgAElEQVSVc+55AOdcut9D8fuaDU9ERGpE0uvw5XVQmh+x+tn0wfxx+fVsmDYw4MCkLqpK8tAIKCh7Y2aHAscCT+/TLhE4osYiExGRg+ccrHwYlo+LXN+gKXcljuaNzPOCjUvqtKokD2uBvsA8/33ZDij7riTSAciombBEROSglRbA4pshcd+RZ096cVtGbh7HN3k/Dzgwqeuqkjw8BfzDzNrg3VVxG7AJ+N8+7c4Hvq/Z8ERE5IAUpMGCwbDji4jVq/J7MCJxAluL25M4TftUSPVUZZGoGWbWCbgVaAt8DdzqnCsua2Nm7YFBaM6DiEj07foWPh0EeUmR67sM4vL3ryUv1DzYuKTeqNI6D865R4BHKqlPR/MdRESib/NsWHR9hRMj/5Z2GaOums2qs7R4sBy4qqzzMKEa53POuSkHEY+IiBwIF4LlE2Dl1IjVRaFGPJgyhjm7fs0oU+IgB6cqPQ+TgHwgl/2vUeoAJQ8iIkEqzvFuw0x+K2J1RklrRm1+kK9yTwg4MKmvqpI8bAS64a31MAt40zkXeQcVEREJVs4GWDAIslZGrF6d352bE8fz+ZQbAw5M6rP99l05544GTgdW4vUqbDOzN8zsCjPTbBsRkWhJfgveT6gwcXgv83Qu2/AoycUdAw5M6ruqTphcCiwF7jGzs4CheLdwPm9mc4G/O+cW1F6YIiKyR2kRfDsW1kyvsMn0bVdz553/5kLNb5BaUO2fKufcAufcaKAr8DfgSuCOgwnCzH5mZt+We2Sb2R1mNsnMUsqVX1jumAfMbL2ZrTGzC8qV9/fL1pvZ2IOJS0Qk5uxOhI/OrDBxyC1txm8TH+QvaVeDEgepJdXektvMzsDrebgcaAXMAZ45mCCcc2uAk/3zNwRSgDeBG4DpzrnH9onhOD+G4/E25frIzMp2/Xwa6AckA0vMbK5zbtXBxCciEhOS34Ivr4fizIjVW4o6clPieNYUdA80LIk/VUoezKwX3h/rK4GOwPvAncBc51xeDcd0HrDBObfZrMKbOwYBs5xzhcAmM1sPnOrXrXfObfTjnuW3VfIgInVXFYYpPsw+lXu23Ml3U4cGGJjEq6qs87AG6AF8DEwE3qjluy2GAq+Uez/GzIbhzbm42zm3C+gMLCrXJtkvA9iyT3nvWoxVRKR25W+Fz6+A9IWR660RU1KG89yOS9j/3fQiNaMqPQ898XbVPAXoBfypkh4BnHMdDjQYM2sCDAQe8IuewbvDo2z9iMeBG4n8f4gj8hwOV8FnjQRGAnTr1u1AQxYRqT3pC+Gzy6FgW8Tq5KL2/C7pft6ccDfjAw5N4ltVkocg96v4DfC1c247QNkzgJn9A3jHf5uMN2GzTBcg1X9dUflenHPPAs8CJCQkREwwRESiwjlY9wwsux1cScQmH2b15p7kO8gqbRVwcCJV2xgryOThKsoNWZhZJ+fcVv/tYH7ctXMu8LKZPYE3YbIn8BVej0RPM+uBN+lyKHB1QLGLiBy8knxYOho2zohYXewaMm3r9RqmkKiq9t0WtcXMWuDdJfHbcsV/MrOT8YYeEsvqnHMrzWw23kTIErxdPkv984wBPgAaAs875yKvniIiEmtyN8OCS2HX1xGr04rbMXrzWJbmHa9ttCWqYiZ58O/aOGyfsusqaT8VCNsBxjn3HvBejQcoIlKbkt+GRcOhaFfk+sNP4+IFo0grOSxyvUiAYiZ5EBGJS6Fi+O5BWP1YhU3+veNCrrvyTb46v0mAgYlUTMmDiEi05CbB51fCzkURqwtDjRmXMprXdvXjuoZKHCR2KHkQEYmGlHfgy+FQlBGxOrmoPbdsfpAV+T0DDkxk/5Q8iIgEKVQM3z0Eqx+tsMlH2f/H3Vvu0mqRErOUPIiIBCVnA3xxNez8KnK9NeTh1GH8I30wrvr7FooERsmDiEgQEl+Gr0ZBSU7E6tSiwxmTdD9vTLiHBwMOTaS6lDyIiNSm4t2w7HcVLvoE8HF2AndtuYvM0tbBxSVyEJQ8iIjUloyvYeFQyFkXsbrENeCxbcP4e/qlGqaQOkXJg4hITXMh+GE6fPeAN0Eygi1FHbkt6V6+yfu5VouUOkfJg4hITcrd4q0UuX1+hU3ezjyTB5PHkBNqGWBgIjVHyYOISE3Z/Ko3KbI4M2J1XqgpE1NG8dquX5M47eKAgxOpOUoeREQOVlEWLP0dJP674jZtf8GAxaPYUNg1uLhEaomSBxGRg5H2GXx5nbcjZgWeSx/EiCtnMe/CZgEGJlJ7lDyIiByIUDGsmAyrHvYmSEawrfhQ7t5yFwt3n8yIhkocpP5Q8iAiUl27N8LCayrc0Arg3cwzeDBlDFmlrQIMTCQYSh5ERKrKOUh8EZbcWuFKkTmlzZmYMoo3Ms/VpEipt5Q8iIhURVEWLLkFNr9SYZMlucdxZ9JdJBcfEWBgIsFT8iAisj/bPoLFN1U4KbLENeDJ7VfxdNoQNkwbGHBwIsFT8iAiUpGiTPjmHtjwXIVNthR15Pake/g679gAAxOJLiUPIiKRJL8NS0ZBfmqFTd7c1ZcJKbew4uEhAQYmEn1KHkREyitIh2W3Vzq3Iae0OeNTRvOfzHMCDEwkdih5EBEpk/QaLBkNhTsqbLJ49wnck3wHn00ewZ8DDE0klih5EBEpzIClt8LmWRW3aXQIDyUO4+WM/to+W+KekgcRiW8p78HiEVCwrcImn2SfQt/r3mBqy25MDTA0kVil5EFE4lNxNnx9N2z4Z4VNMksOYXLqzd6CTy27BRicSGxT8iAi8Wfbx7D4xko3s3o/6zTGp4wmvaRdgIGJ1A1KHkQkfhSke+s2bHqhwibZpS2ZkDKK/2T21fLSIhWIqVk/ZpZoZivM7FszW+qXHWpmH5rZOv+5nV9uZvakma03s+Vm1qvceYb77deZ2fBoXY+IxAgX8hZ6eufnlSYOn+b04vy1T/u3YFpw8YnUMbHY83COc678fVJjgXnOuWlmNtZ/fz/wG6Cn/+gNPAP0NrNDgYlAAuCAZWY21zm3K8iLEJEYkbUKvhoF6Z9V2CS3tBl/2HoTr2RcoN4GkSqIqZ6HCgwCZvqvZwKXlCt/wXkWAW3NrBNwAfChcy7DTxg+BPoHHbSIRFlJPnw3Dv57cqWJw+Ldx9N/3VO8ktEf9TaIVE2s9Tw44H9m5oC/O+eeBTo657YCOOe2mlkHv21nYEu5Y5P9sorKRSRebJvn9TbsXl9hk6ySlkzbdgOzMs5n07QBAQYnUvfFWvJwhnMu1U8QPjSzHyppG+mfCK6S8r0PNhsJjATo1k23YInUCwXp3u2Xif+utNmbu/oydesIduhOCpEDElPJg3Mu1X9OM7M3gVOB7WbWye916ASk+c2Tga7lDu8CpPrlffcp/yTCZz0LPAuQkJAQllyISB3inDcR8pu7oXBnhc02FXZiXMqtLNx9MonTLgowQJH6JWbmPJhZSzNrVfYaOB/4HpgLlN0xMRx4y389Fxjm33XRB8jyhzc+AM43s3b+nRnn+2UiUh9lr4OPfw2Lrq8wcSgKNeIv24fSf+3TLNx9crDxidRDsdTz0BF408zAi+tl59z7ZrYEmG1mI4Ak4Aq//XvAhcB6IA+4AcA5l2FmU4AlfrvJzrmM4C5DRAJRkg+rHoFVf4RQUYXNFu8+ngdTxrChsKt6G0RqSMwkD865jcAvIpTvBM6LUO6AWys41/PA8zUdo4jEiJT3YOkYyN1UYZOskpa0Of3PDP1nR21kJVLDYiZ5EBHZr9wkWHYHJL9ZabO5mWcxJfVmlgy7lk3TAopNJI4oeRCR2FdaCGv+DCsmQ2lehc22FHVkXPJoPt19SoDBicQfJQ8iErucg5R34Ou7Kl2zoSjUiGfTL+WptCH88MhlAQYoEp+UPIhIbMpaBcvuhG3/q7TZwpyTmJB6CxsKu1baTkRqjpIHEYktRbtgxe9h7VPgSitsllbcjj9svYm5mWdpPwqRgCl5EJHYECqFDf+A5eMqXeip1DXghZ0X8cS2a8kJtQwwQBEpo+RBRKJv+3zvLorM5ZU2+2L3Sfw+dSRrCrprzQaRKFLyICLRs3sTfHMvbHm90mZbijryh9QRfJB9Gtr5UiT6lDyISPCKd8OqabD6MQgVVtyuYQv+lHw5z+24hELXRL0NIjFCyYOIBCdUDBuegxWToGB7pU3f3NWXwTe8yH0tOnNfMNGJSBUpeRCR2uectyrktw9AztpKm36b15PJqSP5Ou9YBrfoHFCAIlIdSh5EpHalfQ7f3gc7vqy8WXE7/rjteh6/7zHeMO1FIRLLlDyISO3IWgXfPQjJb1XarDDUiOd2XMLTaUPIDbXgcSUOIjFPyYOI1KzcLbBiImyaCS5UadN3Ms/kT9uGsWDyTYwOKDwROXhKHkSkZhTuhJWPeCtDVnYHBfDl7hN5ZOsNLM8/JqDgRKQmKXkQkYNTkgtr/gKr/gjF2ZU2/SH/J0zbdj2f5CRoSWmROkzJg4gcmNICWPd3WPUwFKRV2jS16HCe2H4tb+w6hxANAwpQRGqLkgcRqZ5QMWx4Hr6fAvkplbdt0o6piYN5YedFFLqmWuRJpJ5Q8iAiVRMqhcSX4Pvfw+6NlTbNDzXl+R0DuXX0szzUpC0PBRSiiARDyYOIVC5UCkmvwveTIXtNpU1LXANezTifv2y/irSSw7i1SduAghSRICl5EJHIQqWweRasnLLfpCHkjLmZZ/Hn7VfzyeSRXBNQiCISHUoeRGRv1UgaAN7POo3p269hTUH32o9NRGKCkgcR8ZQWQeKL3i2X+9l/AmB+9ik8sf1aVuT31ERIkTij5EEk3hXvhg3/gNWP7//uCWBhzklM334NcybcxzkBhCcisUfJg0i8KtwJa/4Ka/8KRRn7bb4w5yT+knY1X+WeEEBwIhLLlDyIxJvcLfDDdNjwrLc65H6UTxo0PCEioORBJH5krYbVf4JNL4Ir2W/zz3N+wZNpVzF7/P2cEUB4IlJ3xMTet2bW1czmm9lqM1tpZrf75ZPMLMXMvvUfF5Y75gEzW29ma8zsgnLl/f2y9WY2NhrXIxJTdiyCBYPh3eNg44xKE4eQM+h6KQPXPcG1m6ZqiEJEIoqVnocS4G7n3Ndm1gpYZmYf+nXTnXOPlW9sZscBQ4HjgSOBj8ysbHu+p4F+QDKwxMzmOudWBXIVIrEiVALJb3kbVqV/tt/mxa4hb+46h7+nX8a8a0Yx98wAYhSROismkgfn3FZgq/86x8xWA50rOWQQMMs5VwhsMrP1wKl+3Xrn3EYAM5vlt1XyIPGhMAM2POdti52XtN/meaGmzNp5ATfe9BRDWnZlSAAhikjdFxPJQ3lm1h34JbAYOAMYY2bDgKV4vRO78BKLReUOS+bHZGPLPuW9K/ickcBIgG7dutXcBYhEQ9YqWPMkbPo3lObtt/muklbM2DGAmTsvJrO0NTe27BpAkCJSX8RU8mBmhwCvA3c457LN7BlgCuD858eBGwGLcLgj8hwOF+mznHPPAs8CJCQkRGwjEtNCpZD6Hqx9ErZ9VKVDUora88/0S5iVcQGrH7mMO2s5RBGpn2ImeTCzxniJw0vOuTcAnHPby9X/A3jHf5sMlP+nUhcg1X9dUblI/VCUBRv/5a3PsJ/dLcusKejG39Mu44n7/sjEBo2ZWMshikj9FhPJg5kZ8Byw2jn3RLnyTv58CIDBwPf+67nAy2b2BN6EyZ7AV3g9Ej3NrAeQgjep8upgrkKklmWtgrX/DzbNqNL6DCFnzMv5P/61YyBf7P4FYDzRoHGthyki9V9MJA94cxuuA1aY2bd+2YPAVWZ2Mt7QQyLwWwDn3Eozm403EbIEuNU5VwpgZmOAD4CGwPPOuZVBXohIjSotgi1vwPpnIG1B1Y5p1Irnt57DzJ0Xs7noSC3sJCI1zpzTcH9CQoJbunRptMMQ+dHuRFj/LGx8DgrSqnTIpsJOzNwxgEl3PAqNW9dufCJSp5nZMudcwoEeHys9DyJSWggpb3vzGVL/SwVzfcN8mtOLf+0YwKc5p+BowCQlDiJSy5Q8iETbrm9hw7+87bCrsEEVQG5pM+bsOo8Xdl7MhsKuGpoQkUApeRCJhsKdkPiy18uw65sqH7auoCsv7fwNr+86jxUPD2F4LYYoIlIRJQ8iQQmVwNYPvP0lUuZCqKhKhxWFGvFB9mm8uPNCXh13P5PMmFSrgYqIVE7Jg0hty1rtJQyJ/4b8rfttXia5qAMv7+zP7F392FHSziu0SOujiYgES8mDSG3IS4WkV72hiYyq38lTEGpCsx6Xc8384/hi90k4Gmg+g4jEHCUPIjWlaBckvQ6bX4btn1DVuyUAvs07htcyfs3bmWex/NoreemMWotSROSgKXkQORjF2ZA8F5Jmw9b3IVRc5UPTi9vyZuY5zMk4j//9/lZOBqbWXqQiIjVGyYNIdRXneOsxJM2G1PchVFj1Q11D5mWfymsZv+a5B8YzskFjb2tXEZE6RMmDSFUUZkDqu95S0an/rVbCALAy/yjmZJzHxN9NpX+z9vSvpTBFRIKg5EGkInkpkPwf2PImpH0C3vYpVbapsBNzM8/m9hsncHybYzm+dqIUEQmckgeRMs5B5gpvSCL5LchYUu1TbC8+lI4nDGPgO91Ynt8TMG5vc2zNxyoiEkVKHiS+leTD9vmQ+g6kvAN5W6p9ivTitvw36wzezfoVS3KPY+Pwgcw9pRZiFRGJEUoeJP7kbPDujNj6AWybB6V51T5FenFb3s86nXezfsVXucezcdpAhtVCqCIisUjJg9R/xbu9OQupfsKwe/0BnSa5qD0fZvfhg6zT9iQM19VspCIidYKSB6l/QsWwc4nXq7D9I9jxZbXWXyhvTUE3Psg6jQ+yT2Nl/k9JnHYxN9RwuCIidY2SB6n7XAgyv/fmLmz7CNI+hZKcAzpVsWvIV7uP5+OcU5mX/X8kFnUmcdpF3FbDIYuI1GVKHqTuCRVDxjeQvgDSFkDaZ1CcecCnyyhpzaFHD+LWT49kQU4vckItSZx2EeNrMGQRkfpEyYPEvuJs2LEY0hfCji+8R0nuAZ+u1DXgu7yefJpzCgt29+K7vJ5sHDaQp0+vwZhFROoxJQ8SW1zIuxti51dekpC+ELJWeOUHYVvxoXyacwpXDriJhkf8ml5ND6UXcGfNRC0iEleUPEj0OAd5yd5iTDv9R8ZSKM466FNnlbTky9yTWLj7F0wZeTtHtP4ZV5rVQNAiIqLkQYIRKoWcdbDrW9j1zY+Pwh01c/5Gh/DprmNYvPsEPt99MnPH30b/Bg21h4SISC1Q8iA1yzko2A5ZK707ILK+954zlx/QYkwVySw5hCW5x7M493gW557I2+PHcHaDRpwN3FdjnyIiIpEoeZADEyqFvM2QvQayf/CfV3tJQ+HOGv+4zYVHsDTvWC47/0pofzpt2xxPP2tAvxr/JBER2R8lD1KxUAnkJXkTGHf7j5wN3vBDzrpqb0tdVbmlzfg+/6d8l38My3KP5e93/o6fND+Cn9TKp4mISHUpeYhXznkLKeWlQG6StyFUXpL/2n/O3QyupFbDKAw1ZnVBD07+RT84LAEO/T9atv45vRs0pHetfrKIiBwoJQ/g/Qu6JB8aNY92JAfHOW9NhMJ0KEj3JiMWpnuP/O1QsBXyUyEv1XuuwTkIVZFd2pJV+T1Ymf9TVuYfxROjbqBp659zcoPGgcYhIiIHp14mD2bWH/gL0BD4p3NuWqUHZH4Ps1tA49bQ7Aho3sl7btYBGh0CjQ+Bhi2hUUvvfaOW0LAZNGwKDZpAg6b+66bQoDFYA6ABWEPvddkjVAqEwJV66xa4Uu8RKoLSQi+JKf9cmu8thlSye5/nHCjK9B+7vNUVi3Z572u5p6AqCkON2FDYlTUFP2FtwU+4/8rLoO0JtG7ZnT5m9Il2gCIiclDqXfJgZg2Bp4F+QDKwxMzmOudW7ffg4mzvkbO2lqOsJxq35evMjmwo7MrGws7cP2QQtD6Wpq2O5rgGjTgu2vGJiEitqHfJA3AqsN45txHAzGYBg4D9Jw8SZmdJa5IKO7G56Ag2F3Xi9kH94ZCfQutjoGl7epnRK9pBiohIoOpj8tAZ2FLufTKEz70zs5HASIBTegQTWMxp0ITkgrakFh9OalF7LjnjNGjZDVp02/N8WJM2HAb8MtqxiohIzKiPyUOkNYhdWIFzzwLPAiQc3cRh/hyEOi4/1JSdJa3JKGnjPUpbc2mfk735G807QfMjf3xu0o4uZnSJdtAiIlKn1MfkIRnoWu59FyC10iPangRDv/IWNyrYBvlbIX8bFO30JyiWn6zov95rgmPRj69dyY+TIQmVmxgZ8idQlk2ibPjjxMoGTX6ccFn+uWHzcpM0D9l7wmaTttC4LTRp571u4r1u3rAZXfyLFhERqQ31MXlYAvQ0sx5ACjAUuHq/R1kDaNbee7Q9sZZDFBERqbvqXfLgnCsxszHAB3i3aj7vnFsZ5bBERETqjXqXPAA4594D3ot2HCIiIvVRg2gHICIiInWLkgcRERGpFiUPIiIiUi1KHkRERKRalDyIiIhItSh5EBERkWpR8iAiIiLVYs6FbfsQd8wsB1gT7Tii6HBgR7SDiJJ4vnbQ9ev64/f64/naAX7mnGt1oAfXy0WiDsAa51xCtIOIFjNbGq/XH8/XDrp+XX/8Xn88Xzt4138wx2vYQkRERKpFyYOIiIhUi5IHz7PRDiDK4vn64/naQdev649f8XztcJDXrwmTIiIiUi3qeRAREZFqievkwcz6m9kaM1tvZmOjHU9tM7OuZjbfzFab2Uozu90vP9TMPjSzdf5zu2jHWpvMrKGZfWNm7/jve5jZYv/6XzWzJtGOsbaYWVszm2NmP/g/B6fFy/dvZnf6P/ffm9krZtasPn/3Zva8maWZ2fflyiJ+1+Z50v9duNzMekUv8ppRwfU/6v/sLzezN82sbbm6B/zrX2NmF0Qn6poT6frL1d1jZs7MDvffV/v7j9vkwcwaAk8DvwGOA64ys+OiG1WtKwHuds4dC/QBbvWveSwwzznXE5jnv6/PbgdWl3v/R2C6f/27gBFRiSoYfwHed879HPgF3n+Hev/9m1ln4DYgwTl3AtAQGEr9/u5nAP33Kavou/4N0NN/jASeCSjG2jSD8Ov/EDjBOXcSsBZ4AMD/PTgUON4/5v/5fyPqshmEXz9m1hXoBySVK6729x+3yQNwKrDeObfROVcEzAIGRTmmWuWc2+qc+9p/nYP3h6Mz3nXP9JvNBC6JToS1z8y6ABcB//TfG3AuMMdvUm+v38xaA2cBzwE454qcc5nEz/ffCGhuZo2AFsBW6vF375xbAGTsU1zRdz0IeMF5FgFtzaxTMJHWjkjX75z7n3OuxH+7COjivx4EzHLOFTrnNgHr8f5G1FkVfP8A04H7gPITHqv9/cdz8tAZ2FLufbJfFhfMrDvwS2Ax0NE5txW8BAPoEL3Iat2f8f7HCfnvDwMyy/1Cqc8/B0cB6cC//GGbf5pZS+Lg+3fOpQCP4f1rayuQBSwjfr77MhV91/H4+/BG4L/+67i4fjMbCKQ4577bp6ra1x/PyYNFKIuLW0/M7BDgdeAO51x2tOMJipldDKQ555aVL47QtL7+HDQCegHPOOd+CeRSD4coIvHH9gcBPYAjgZZ4XbX7qq/f/f7E0/8HmNlDeMO4L5UVRWhWr67fzFoADwETIlVHKKv0+uM5eUgGupZ73wVIjVIsgTGzxniJw0vOuTf84u1lXVT+c1q04qtlZwADzSwRb5jqXLyeiLZ+VzbU75+DZCDZObfYfz8HL5mIh+//18Am51y6c64YeAM4nfj57stU9F3Hze9DMxsOXAxc435cqyAerv+neMnzd/7vwC7A12Z2BAdw/fGcPCwBevqzrZvgTZaZG+WYapU/vv8csNo590S5qrnAcP/1cOCtoGMLgnPuAedcF+dcd7zv+2Pn3DXAfOByv1l9vv5twBYz+5lfdB6wivj4/pOAPmbWwv//oOza4+K7L6ei73ouMMyfdd8HyCob3qhPzKw/cD8w0DmXV65qLjDUzJqaWQ+8iYNfRSPG2uKcW+Gc6+Cc6+7/DkwGevm/F6r//Tvn4vYBXIg343YD8FC04wngen+F1xW1HPjWf1yIN+4/D1jnPx8a7VgD+G/RF3jHf30U3i+K9cBrQNNox1eL130ysNT/GfgP0C5evn/g98APwPfAv4Gm9fm7B17Bm99R7P+hGFHRd43Xbf20/7twBd5dKVG/hlq4/vV4Y/tlv//+Vq79Q/71rwF+E+34a+P696lPBA4/0O9fK0yKiIhItcTzsIWIiIgcACUPIiIiUi1KHkRERKRalDyIiIhItSh5EBERkWpR8iASBf791Jv8ne2OPoDjO5jZJH+Z8Vrhn3/HARx3iH9d11fzuBlmtrS6n3cwzOxtM5tYA+fp7l/zxdU8LsHMdppZm4ONQSRISh5EouM0oLv/eugBHN8BmFjuHFJNZtYbOAf4aw2cbived/p5dQ5yzi3FW2/gzhqIQSQwSh5EouMqvL0lFvuvJXi3AW855yLtPFgtztuNcZHzdimtrn8Bo8otky0S85Q8iATMzBoCV+AtCfs8cJyZnRSh3U/M7BUz22FmeWa23Myu9ocqVvjN5vvd5c4/5nr//SH7nCvRzB4r9/4iM/vQzNLMLNvMFpnZ+Qd4PZeZ2Vozy7f/396ZhlpVRXH8948mNc3KssTMKKSsPpglmUTph0qwQatXYSMN0JxNVFbkUJaNRiJUWhZFpTTYSFq9DDM1Ci0lm0DTsLK5UJ/m6sPaR8+73nufR/L6LqwfPM45+66zznrny1l77bXXkmYCB1WQu1jSQklrJC2RdFMLeveRNEnSd0n3V5JGp3Lymcw8SU+WuXeypE+r6G4PDGZjO+5svFHSVEkXpmWlvyU9k8oW95E0N401SuqWu2+TZYvsnUsaJmmZpN8kPS+pY2yrQ9sAAAVLSURBVIk504DdgROqvY8gaE2E8xAEtWcA0BlvzjUVLx/bLPogaS9gNnAkcANwEt6XZF88RD40iV6Bh8v7FrRhf+A14FzgNOAj4C1J/YookXQ48AIwHxiCfwhfLCN3IzABL4k9KJ2PknRlFfWdgF+B64ATgfuAC2m+zPAEcEbeWUrnp+Ez+kocDbTB/+9SjsL7PlyFt29vSM98HBgHnIOXtX6siv6MBryPxqV4T4VBwN15AfPOtgvx5l1BUBdEmCwIas/ZwO/A22bWJGk63pTnVttYL34YsCvQ2zY2qHk3UyBpQTpdZGYfFzXAzB7N6doObxB1CF7/f1YBVTfj/WEaku1vSdoJGJ3T3wHPzxhtZiPS8HR5i+DbJE0ws3/L2Pg57jhlembhSz2TJF1lZk14/f4H8UhO5iw0ADsAz1Wxuzew0sx+LPPbLsApZvZHeu5xwCXAsWY2M411AcZLamvNGyyVshY41czWpft64jkul5fIzQf6VNETBK2KiDwEQQ1JH9bBwMvp4wf+AeyOz3gzBuDOxVbpbCipawrtLwfW4R+544EeBVX1AaZZ8yY5L5XI9AXaAVMkbZ/9Ae/hEZiuFWyUpGslLZK0Ktn4LN7QqhtsmLVPBS7I3XpBsumXKnbvDVTaSfJJ5jgkvgGaaJ4M+U06dqnyDID3M8chsQjYK7/0kliZbAqCuiAiD0FQWwYCHYE3c2vfjcAaPCIxO43tgbeN/99JkYZpQHvgDvxD+A8wEt/FUYS9gZ9KxkqvO6Xjwgo69gWWlBm/FrgfuAf4APgNX8YZD+yck5sINEo6IF0fg3eLrcbO+DsvR2nSYxPwl5mtLxnL9FSjnC4BO+Z0kGxpSVcQtBrCeQiC2pLlNkwp81uDpGEphP8LsM8W6F+djqUz291y5wcCvfC2w29ng5LabMHzVrCpw1F6ne1mGASUWyZYXEH3GcAUMxues7FnqZCZzZT0NZ6nIOAH4J0W7P4Vd+JaCx3Z+J6CoNUTzkMQ1IiUyDcIX6YoTbbrha/d9wdm4PkNV0vqXGFdvtLMd1k6HkzKXUj1DDrkZDInYcPMW9J+QD9gAcWYB5ws6Zbc0sWQEpnZwCqgi5m9UUB3GzaNDgwtJ4jvWsnyCJ4ul0NRwmKgi6SdzKxSBKKWdMdzR4KgLgjnIQhqxylAW2Ccmc3J/5CSAYfjkYkZwEPAecCHku4CvscdgnZmNhZYin+Qz5f0B7A2FRyaCywHHpF0O74F8Cbgz9zjvsSdjAeSTHtgRLqvKPfitSpelDQROBRPutyAmf0u6U5gXHJSZuL5Vj2A/mY2uILu6bgDNQf4FnccKlXjnIwnaW4PPLUZds/CkyoPA2pa1bICR+DvMgjqgkiYDILacTbwdanjAGBma/EtjkPSbPhnPBLwGfAw8Dq+3W9pkl+N7wDojecDzEvjTXhC5no8kfB64DI8XyB71ho8OrAuyYwCxiQ9hUgOy1l45OQV4FTgzDJyY5P9A4FX8ejLUODDKupHJrnR6diEF3YqZ8cK3ImZZWaVlkHy8l8BXyR7timSegF7smmiaRC0WtQ8SToIgqD+kLQ7Hjm50swmbuY9w4CLzOzQrWpcy3aMAY40s6jzENQN4TwEQVC3pEqRPYFr8CJL3Vuou5C/tx3wHTDUzGZsPStbtGEJcLqZNW4LG4JgS4ichyAI6pneeIGrJcB5m+s4AJjZP5LOx2tQbCu6ASPDcQjqjYg8BEEQBEFQiEiYDIIgCIKgEOE8BEEQBEFQiHAegiAIgiAoRDgPQRAEQRAUIpyHIAiCIAgKEc5DEARBEASF+A/w099rKdFX5gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8,5))\n",
    "ax.scatter(sorted_pred_df_dum.target, sorted_pred_df_dum.mse, label='MSE', s=3)\n",
    "ax.set_xlabel(\"Actual delay (min)\", fontsize=15)\n",
    "ax.set_ylabel(\"MSE\", fontsize=15)\n",
    "ax.set_xlim((0,140));\n",
    "ax.plot(delays, mean_mse_delay_dum, label='Mean MSE', color='orange', linewidth=5)\n",
    "ax.legend(fontsize=15);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From this simple plot we can see that this dummy model : \n",
    "* Performs poorly as delay increases\n",
    "* Performs best for delays between 0 and 20 min, which may be an interesting feature as most delays are small delays"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sizes = {} # clé : coordonnées ; valeur : nombre de points à ces coordonnées\n",
    "for (yt, yp) in zip(list(y_test), list(y_pred_dum)):\n",
    "    if (yt, yp) in sizes:\n",
    "        sizes[(yt, yp)] += 0.01\n",
    "    else:\n",
    "        sizes[(yt, yp)] = 0.01\n",
    "\n",
    "keys = sizes.keys()\n",
    "fig, ax = plt.subplots(figsize=(8,5))\n",
    "plt.scatter([k[0] for k in keys], # vraie valeur (abscisse)\n",
    "            [k[1] for k in keys], # valeur predite (ordonnee)\n",
    "            s=[sizes[k] for k in keys], # taille du marqueur\n",
    "            color='coral')\n",
    "\n",
    "ax.set_xlabel(\"Actual delay (min)\", fontsize=15)\n",
    "ax.set_ylabel(\"Predicted delay (min)\", fontsize=15)\n",
    "ax.set_xlim((0,140));"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Save Memory"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "del y_pred_dum, pred_df_dum, sorted_pred_df_dum"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4.3) Baseline : Linear regression with selected features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This will be our baseline model : a simple linear regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "lr = linear_model.LinearRegression()\n",
    "# Training the model with training set\n",
    "lr.fit(X_train, y_train);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Linear regression on training set : RMSE : 20.79, MAE : 12.45, R² : 0.03\n",
      "Linear regression on testing set : RMSE : 19.73, MAE : 12.05, R² : 0.01\n"
     ]
    }
   ],
   "source": [
    "display_metrics(\"Linear regression on training set\", lr, X_train, y_train)\n",
    "display_metrics(\"Linear regression on testing set\", lr, X_test, y_test);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The score may be poor, the RMSE is close on both training and testing sets, thus the model might be lausy representation of the data but consistent anyhow"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looking closely at our prediction error behaviour : mean squared error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    1.355448e+06\n",
       "mean     3.892597e+02\n",
       "std      1.495334e+03\n",
       "min      7.854280e-11\n",
       "25%      2.727532e+01\n",
       "50%      7.244452e+01\n",
       "75%      1.491100e+02\n",
       "max      1.925446e+04\n",
       "Name: mse, dtype: float64"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test_predict = lr.predict(X_test)\n",
    "pred_df, mse_stats_df = pred_mse(y_test, y_test_predict)\n",
    "mse_stats_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So the max mean squared error is quite high ! Some predictions are quite bad. The mean and the median however (respectively 6e-3 and 80 are not that high values which may be a good point)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Getting mean RMSE for 2-min delay windows\n",
    "delays = []\n",
    "mean_mse_delay = []\n",
    "delay = 0\n",
    "for i in range(70) :\n",
    "    mean_mse_delay.append(pred_df.loc[:, 'mse'][pred_df.loc[:, 'target'].between(delay, delay+2)].mean())\n",
    "    delays.append(delay)\n",
    "    delay += 2\n",
    "sorted_pred_df = pred_df.sort_values(by=['mse'], ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8,5))\n",
    "ax.scatter(sorted_pred_df.target, sorted_pred_df.mse, label='MSE', s=3)\n",
    "ax.set_xlabel(\"Actual delay (min)\", fontsize=15)\n",
    "ax.set_ylabel(\"MSE\", fontsize=15)\n",
    "ax.set_xlim((0,140));\n",
    "ax.plot(delays, mean_mse_delay, label='Mean MSE', color='orange', linewidth=5)\n",
    "ax.legend(fontsize=15);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Chocking : We find the same behaviour as our dummy mean-predicting model. This Linear Regression model seems to be empty."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAg8AAAFFCAYAAABrDg41AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3XlYlFX7wPHvwQhZxAV3LHd90xZ/SWpmZahlZrjkgppL+LqmZRlmrryWaW6QuWuKFYiKmfue+4qWpbnvueMGyCbL+f0xw8gwA6LCAHJ/rmsueO5zZuaeQZmb5zmL0lojhBBCCJFZdjmdgBBCCCHyFikehBBCCPFQpHgQQgghxEOR4kEIIYQQD0WKByGEEEI8FCkehBBCCPFQpHgQQgghxEOR4kEIIYQQD0WKByGEEEI8lKdyOoHcoHjx4rpChQo5nYYQQghhEwcOHLihtS7xqPe3afGglJoLNAeua62fN8YWAtWNXYoAd7TWtZRSFYCjwHFj2x6tdW/jfWoDgYAjsBr4VGutlVLFgIVABeAc0E5rfftBeVWoUIH9+/dnwSsUQgghcj+l1PnHub+tL1sEAk1TB7TW7bXWtbTWtYAlwK+pmk+ntKUUDkbTgZ5AVeMt5TEHA5u01lWBTcZjIYQQQmQhmxYPWuttwC1rbUopBbQDFmT0GEqpMoCr1nq3Nuzq9RPQ0tjcAphv/H5+qrgQQgghskhuGjD5OnBNa30yVayiUupPpdRWpdTrxpg7cDFVn4vGGEAprfUVAOPXkuk9mVKqp1Jqv1Jqf3h4eNa9CiGEEOIJl5uKhw6Yn3W4Ajyrtf4/4HMgWCnlCigr933ofcW11rO01h5aa48SJR55zIgQQgiR7+SK2RZKqaeA1kDtlJjWOh6IN35/QCl1GqiG4UxDuVR3LwdcNn5/TSlVRmt9xXh547ot8hdCCCHyk9xy5qExcExrbbocoZQqoZQqYPy+EoaBkWeMlyOilFL1jOMkugDLjHdbDnQ1ft81VVwIIYQQWcTWUzUXAA2B4kqpi8BIrfWPgDeWAyXfAEYppRKBJKC31jplsGUf7k/VXGO8AYwFFimlugMXgLZZkXdkZCTXr18nISEhKx5OCJHH2dvbU7JkSVxdXXM6FSHSlxQHBQpmy0Mrw4SF/M3Dw0Ont85DZGQk165dw93dHUdHRwwnO4QQ+ZXWmtjYWC5dukSpUqWkgBC5U/QF2PA6vDgKKnW1aFZKHdBaezzqw+eWyxa51vXr13F3d8fJyUkKByEESimcnJxwd3fn+nUZViVyoYQo2Po+xFyAPd3g4BDQyVn6FFI8PEBCQgKOjo45nYYQIpdxdHSUS5ki90lOgp0d4c7f92NHxsCOtpAYk2VPI8VDJsgZByFEWvJ7QeRKBwfB5ZWW8Rt74N6dLHsaKR6EEEKIJ8GpWXBskmW8gBO8uQKcymbZU0nxkA/4+fmhlKJq1apW26tUqYJSCj8/P9sm9gDdunVDKUWTJk0s2mJjYylUqBBKKQIDA03xhIQEJk2axPPPP4+TkxPFixenbt26jB071tTn3LlzKKWs3ho3bpxhTkoppkyZYjpu2LAhbdq0efwXK7JE2p+PEPnG1U0Q9rGVBgX1f4FiL2fp0+WKRaJE9itYsCBnz55l//79eHjcH2AbFhbG+fPnKVgwe6bzPC4XFxc2b97MtWvXKFWqlCm+cqWV03JAv379CAoKYujQodStW5c7d+6wZ88eVqxYweDB5vukTZgwgddee80sVrhw4YfKb9q0adjb2z/UfUT22b17NxUrVszpNISwrYhjsP0D0IkWTReLj6XcM62y/CmleMgnnJ2defnllwkJCTErHkJCQvD09OTAgQM5mF36qlevTlRUFIsXL6Zfv36meEhICF5eXgQHB5tiMTExzJs3j9GjR+Pr62uKt27dGmtTkqtXr069evUeK78aNWo81v2zQ1xcXKaLwYfp+6hiY2NtNuj4cX+eQuQ58Tdha3NIiLBomrvlIz5d4Muvv4KVE7iPRS5b5CPe3t4sWrTI9EGqtWbRokV4e3tb7b9jxw7efPNNnJyccHNzo0ePHkRFRZnar1y5go+PD5UqVcLR0ZFq1aoxbNgw7t27Z+qTcolg0aJF9OrVi8KFC1OuXDlGjhxJcnLmpg61b9+ekJAQ03FUVBSrV6+2yDs6OpqEhARKly5t8RjZNbgt7WULPz8/ihcvzp9//km9evVwcnLi//7v/9i+fbvFfefMmUPNmjVxcHCgfPnyjBs3zqx99+7deHl5UbZsWZydnalVqxZBQUFmfQIDA1FKsW/fPho2bIijoyPjx4+3mmtGfePi4hg0aBDPPPMMDg4OvPTSS6xevdrs/vHx8fTp04ciRYrg5uaGr68vAQEBZu/tli1bUEqxbt06vLy8cHFxMRV9ycnJjB07lipVquDg4EC1atWYP3++2XPs2LGD119/HVdXV1xdXalVqxaLFy82tS9fvpzatWvj7OxM0aJFqVu3Llu3bjW1W7tsMWXKFKpWrYqDgwNVqlTB39/frP1hfmZC5CpJ8bC9Ndw9bdG09egb9J47g7t3Fe+9B6l+hWYJKR4eklK55/awWrduzbVr19ixYwcA27dvJzw8nFatLE9p7dy5k0aNGlG6dGlCQ0MJCAhg9erVfPTRR6Y+N27coFixYkyaNIm1a9fi6+vLvHnz6N+/v8XjDRo0CBcXF0JDQ/nwww8ZNWoUoaGhmcq7Q4cO7Nq1iwsXLgCwdOlSihYtyptvvmnWr0SJEjzzzDP4+fnx66+/mhU61iQnJ5OYmGh2y2xBk5GYmBi6du1Kr169WLJkCQ4ODrRq1YqYmPvTpMaPH0+fPn1o2bIlK1eupE+fPgwfPtzsg+/8+fO89tprzJkzhxUrVvDBBx/w0UcfsWCB5a71HTp0oHnz5qxevZrmzZtnmJ+1vm3atCEwMJAhQ4awYsUKXnnlFby8vDh48KDpfoMGDSIwMJCRI0cSFBTEhQsXmDhxotXn6N69Oy+99BLLly+ne/fuAPTv359vvvmGnj17smrVKlq1aoWPj4/pElRkZCTNmzenUqVKLFmyhNDQUDp37sydO4YR4qdPn6ZNmzZ4enqyYsUKgoKCaN68Obdu3bKaA8Ds2bPp378/Xl5erFixgrZt2zJw4ECzMTCZ/ZkJkatoDft6wfVtFk0nr1ahdcCvJCQ9DUBCAnTqBCdPWnR9nOfX+f5Wu3ZtnZ4jR46YHRt+YrnjllkjR47Ubm5uWmutvby8dN++fbXWWvfp00e3aNFCa621m5ubHjlypOk+DRo00A0bNjR7nE2bNmlAHzp0yOrzJCQk6KCgIO3g4KDj4+O11lqfPXtWA7pz585mfV966SXdvn37DPPu2rWrTvnZvPjii3rcuHFaa63fffdd/emnn+qoqCgN6Hnz5pnlWKJECQ1oOzs7Xbt2bT1+/HhTPqlzsnZL/R5YA+gffvjBdPzmm2/qDz74wHQ8cuRIDehNmzaZYn/++acG9Jo1a7TWWkdERGhnZ2ft5+dn9tjDhw/XpUqV0omJiRbPm5ycrBMSEnTPnj31W2+9ZYrPmzdPAzogICDDvDPqu3HjRg3oLVu2mMVff/113aZNG6211jdu3NAFCxY0/QxScqpRo4Ym1T/GzZs3a0APGDDA7LFOnjyplVI6MDDQLN65c2ft4eGhtdY6LCxMAzoyMtJq/osXL9bFihXL8DWm/vkkJSXpsmXL6m7dupn16dOnj3Z1ddWxsbFa68z9zNKT9veDEDZz6Butg7C43ZpVRFcrc8zi8yLtrwhgv36Mz00585DPeHt7ExoaSnx8PKGhoVYvWcTExLB7927atWtn9ld5gwYNsLe3N42P0FoTEBBAjRo1cHR0xN7enk6dOhEfH286S5Di7bffNjuuUaMGFy9eJLO8vb0JCQnh1q1bbNy4Md1LLZ6enpw+fZoFCxbg4+PDzZs38fX1xdPT0+Ksgr+/P2FhYWa3nj17Zjqn9Njb29OwYUPTccq4iJTXu3v3bqKjo2nbtq3Z++vp6cm1a9dM/W7fvs0nn3xC+fLlsbe3x97enlmzZnHixAmL53zvvfcynV/avhs3bqR06dK89tprZvk0atSIlGXbDx06RFxcHF5eXqb7KaV4//33M/UcmzZtws7OjlatWlk8x8GDB0lKSqJy5cq4uLjQsWNHli1bZjrjkOKFF14gIiKCrl27sn79eqKjozN8nRcvXuTy5cu0bWu+xU379u2JjIzk0KFDptiDfmZC5CrnFsDfwyzCCYlP0eb7UE5cqW4WHzwYPv00a1OQ4iGf8fLy4u7duwwdOpTo6Girv/xv375NUlISffv2NX1o2dvb4+DgQEJCAv/++y8AAQEBDBw4kFatWrFs2TL27dvH1KlTAcM19NSKFClidvz0009b9MmIt7c3f/zxB99++y3u7u4ZDowrVKgQ3t7ezJ49mzNnzjB8+HB27tzJihUrzPpVqVIFDw8Ps1vZso8/D9rV1RU7u/v/tZ5+2nDqMOX13rhxA4CaNWuavb9vvfUWgOn97datGwsXLsTX15f169cTFhaGj4+P1fct9UyUB0nb98aNG1y9etUsF3t7e/z8/Ey5XL16FTBcGkot7XFGz5GUlEThwoXNnqNbt24kJiZy5coVihYtyvr160lISKBdu3aUKFGC9957jzNnzgCGAa7Lli3jzJkzNGvWjOLFi9OxY0fCw8Ot5nDlyhWruaQcp77c8aCfmRC5RvhOw5LTVvQNnMbv/zQyi330EXz7bdanIbMt8hlnZ2eaN2+Ov78/bdu2xdnZ2aJPkSJFTOs+NGvWzKI95QN28eLFtG3bltGjR5vajhw5ki15V6xYkTp16uDv7282k+JBlFL4+vry9ddfc+zYMVq0aJEt+T2MYsWKAYbpptY+9KtXr05cXByrVq1iypQp9O7d29SW3piMhxkQmrZvsWLFcHd357fffkv3PimDUMPDw035pxxn9jmeeuopdu7cafYhnaJkyZIAvPrqq6xdu5bY2Fg2btzI559/TseOHdmzZw9gOKPx3nvvERERwapVqxgwYAD9+/c3G1CbokyZMgAW+09cu3bNlJMQeUrUKdjWApLvWTR9t2IQczb3MIs1bw6zZj3aGLkHkeLhIeknYBPSPn36EB8fb/ahlJqzszP16tXj+PHjjBgxIt3HiY2NxcHBwSyWdjZAVho4cCDBwcF06dLFantCQgLR0dEWZzlOGkcJPcxf59np1VdfxdHRkcuXL6d7uSEiIoKkpCSz9zcqKorly5dn+cyRRo0aMXHiRFxcXPjPf/5jtc8LL7xAwYIFWbZsGYMGDQIMl63Sns1Jj6enJ0lJSURERFhd9CstR0dH3n//fQ4fPsyYMWMs2gsXLkzHjh3ZunUru3fvtvoY5cqVo2zZsixevJh3333XFF+0aBGurq688MILmcpdiFwh/hZsec8wNTON0H0f8NVC8/8nr74KCxfCU9n0KS/FQz7UsGFDs+u71owbN45GjRphZ2dHmzZtKFSoEBcuXGDVqlWMHj2aatWq0aRJEyZPnkzdunWpXLkyQUFBnDp1KtvybteuHe3atUu3PSIigmrVqtG1a1feeustChcuzPHjxxkzZgzu7u4Ws0qOHz9O8eLFzWIFCxakVq1a2ZJ/iiJFiuDn58enn37K+fPneeONN0hOTubEiRNs3ryZpUuXUrhwYV555RVGjRplOqU+duxYChcuTGRkZJbm06RJE9555x2aNGnCl19+Sc2aNYmMjOTgwYPExcUxZswY01TdkSNHYm9vz3PPPce8efOIjIzMVDFTvXp1evfujbe3N4MGDcLDw4O4uDj++ecfTpw4wZw5c1i1ahVz586lZcuWPPvss1y6dImZM2fi6ekJwMyZM9m9ezdNmzalbNmynDx5ksWLF6dbTNrZ2eHn50evXr1wc3OjSZMmbN26lenTp/Ptt9/m2oXRhLCQdM8wJTPKcrzT3lN16DL9J7S+f0bvuedg5Upwcsq+lKR4EFY1aNCAbdu2MXLkSDp37kxSUhLly5enadOmpr/gR4wYQXh4OMOGGQbutG7dmsmTJ6c7iC67ubq6MmjQIFavXk1wcDCRkZG4u7vzzjvvMGzYMIvVI7/44guLx6hcuXK2FkApBg0aRNmyZfH392fixIkULFiQatWq0b59e1Of4OBgevbsSZcuXXBzc6Nfv37ExMRk+fLLSil+/fVXvv32WwICArhw4QLFihWjVq1aZtNux40bR0JCAn5+ftjZ2dG5c2e6d+9OQEBApp5n6tSpVKtWjdmzZzNixAhcXV2pUaOGaSpnyjLpQ4YM4fr165QoUYLmzZvzrfGC7Ysvvsjy5cv5/PPPuXXrFmXKlKFHjx6MGjUq3efs0aMH8fHxBAQE8P3331OuXDkmTpzIZ5999hjvmBA2pDXs6wHXt1o0nQsvj9fE5cTeu18llCsH69ZBdl+VU/pJOA//mDw8PHTKqPK0jh49ynPPPWfjjITIGxo3bkxCQoLZQk35ifx+ENnur2Hwz2iLcESMK6/67ebopfur3BYtCjt2QGYWvlVKHdBaezy4p3Vy5kEIkSmbN29m7969vPzyyyQkJLBw4UI2bdpktgKkECILnZxutXBISHyKDwKWmBUOjo6GSxW2WjFfigchRKa4uLjw22+/MWbMGOLi4qhatSqBgYGyq6gQ2eHf32B/P6tNfeZNZ9M/93cALlDAMDiyfn1bJSfFgxAik1555RXTlEkhRDYK3wW7OoC2nJr99dJh/Ljlv2axmTPB1kPNZJEoIYQQIreIOAZb34ckywXK5m3txohQ8wHC33wDxjHHNiXFgxBCCJEbxFyGze/APcvN3lYffJeeP84C7k+N/vhjGDLEhvmlIsWDEEIIkdMSImFLM4i5YNEUdtqDdpMXkZhkb4q1aQPff589q0dmhhQPQgghRE5KToDtbeHOXxZNp65W5r0Jq4iOdzHFGjaEn382DJTMKVI8CCGEEDlFawjrC1fXWzRdjyhB03FrCY8saYq98AL89hvk9AKpUjwIIYQQOeXIGDg9xyIcHedE8wkrOX2tiinm7g6rV0OaxXJzhBQPQgghRE44Fwx/DbUIJyXb0X7KQsLO1DHFXF1hzRrD8tO5gU2LB6XUXKXUdaXU4VQxP6XUJaXUQeOtWaq2r5RSp5RSx5VS76SKNzXGTimlBqeKV1RK7VVKnVRKLVRKPW27V5d7+fn5oZSiatWqVttT9hTw8/OzbWIP0K1bN5RSVndhjI2NpVChQiilCAwMNMUTEhKYNGkSzz//PE5OThQvXpy6desyduxYU59z586hlLJ6a9y4scVzPcgXX3xBhQoVHuUlCiHyq+vbYM9HVpv6z/+BVX82Nx0/9RQsWWK4ZJFb2HqRqEBgCvBTmri/1npC6oBSqgbgDdQEygIblVLVjM1TgSbARSBMKbVca30E+M74WCFKqRlAd2B6dr2YvKRgwYKcPXuW/fv34+FxfznzsLAwzp8/n2t3GHRxcWHz5s1cu3bNbEvtlStXWu3fr18/goKCGDp0KHXr1uXOnTvs2bOHFStWMHjwYLO+EyZM4LXXXjOLpd08SwghslzkcdjWEpLvWTSNX/kF0zf2NYvNmQOP8HdNtrJp8aC13qaUqpDJ7i2AEK11PHBWKXUKSDmHc0prfQZAKRUCtFBKHQU8gY7GPvMBP6R4AMDZ2ZmXX36ZkJAQs+IhJCQET09PDhw4kIPZpa969epERUWxePFi+vW7v1RrSEgIXl5eBAcHm2IxMTHMmzeP0aNH4+vra4q3bt0aaxvAVa9enXr16mXvCxBCiNTirhumZN67bdG0eG8bvgz5ziw2ahR07Wqr5DIvt4x56KeU+tt4WaOoMeYO/Juqz0VjLL24G3BHa52YJm6VUqqnUmq/Ump/eHh4Vr2OXM3b25tFixaZPki11ixatAhvb2+r/Xfs2MGbb76Jk5MTbm5u9OjRg6ioKFP7lStX8PHxoVKlSjg6OlKtWjWGDRvGvXv3q+mUSwSLFi2iV69eFC5cmHLlyjFy5EiSky2XXrWmffv2hISEmI6joqJYvXq1Rd7R0dEkJCRQunRpi8dQWTQZ+s6dO3Ts2BFnZ2fKlCnD6NGWm9b4+flRvHhxqzmk3k67QoUKfPHFF4wdO5YyZcpQuHBhBg4ciNaa1atXU7NmTQoVKkTLli25ffv+L5otW7aglGLTpk20aNECZ2dnqlatyvr160lKSsLX15fixYvj7u7OpEmTTPdbtWoVdnZ2nD171iyvs2fPYmdnx/Lly9N93Uop/P39GThwIG5ubhQvXpwJEwwnC+fPn0+lSpUoUqQIPj4+xMWZr4x34cIFvL29KVasGE5OTrzzzjscP37crM/gwYN54YUXcHFxoVy5cnTq1ImrV6+a9Ul5v/z9/SlXrhxFixbF29ubO3fupJu3ELlKYrRh9ci7Zyyadp14lS7Tf0Lr+x/LPj4wbJgtE8y83LC3xXTga0Abv04EfEi9jNZ9GusFj86gv1Va61nALDBsyZ3pbINzaEUOazo+3HbqrVu3pk+fPuzYsYPXX3+d7du3Ex4eTqtWrcz+UgfYuXMnjRo1omXLloSGhnLz5k0GDx7M7du3CQ0NBeDGjRsUK1aMSZMmUbRoUU6cOIGfnx/h4eHMnDnT7PEGDRrEBx98QGhoKJs2bWLUqFHUrFmTdu3aPTDvDh068M0333DhwgWeffZZli5dStGiRXnzzTfN+pUoUYJnnnkGPz8/nJ2dadKkCYUKFUr3cZOTk0lMTDSL2dnZYWeXfk390UcfsWXLFgICAihdujQTJkzg9OnTPPXUo/1XCgkJoU6dOsybN48DBw4wbNgwkpOT2bZtG19//TWxsbH069ePr776ihkzZpjdt1evXvTq1YuPP/6YcePG0aZNGzp16oTWmuDgYFatWsXAgQOpX78+9erVo2nTppQtW5b58+ebjW8JDAykRIkSNGvWjIxMnDiR9957jwULFrBy5Up8fX25fv06YWFhTJ48mQsXLvDZZ59RrVo10yWiW7du0aBBA9zc3JgxYwZOTk6MHTuWxo0bc+LECRwdHQG4fv06Q4YMoWzZsoSHhzNx4kQ8PT05dOgQBVJNZl+0aBEvvvgis2bN4uLFi3z++ecMGTKEadOmPdL7L4TNpKzlcHOfRdOpq5VpMWkZcQmOptg778CMGTm3CNQDaa1tegMqAIcf1AZ8BXyVqm0d8Krxti5V/CvjTQE3gKeMcbN+Gd1q166t03PkyBHzQBC555ZJI0eO1G5ublprrb28vHTfvn211lr36dNHt2jRQmuttZubmx45cqTpPg0aNNANGzY0e5xNmzZpQB86dMjq8yQkJOigoCDt4OCg4+PjtdZanz17VgO6c+fOZn1feukl3b59+wzz7tq1q0752bz44ot63LhxWmut3333Xf3pp5/qqKgoDeh58+aZ5ViiRAkNaDs7O127dm09fvx4Uz6pc7J2S/0epHX48GEN6JCQEFMsKipKFy1aVJcvX94US/1+pwboH374wXRcvnx5XblyZZ2YmGiKvfLKK7pAgQL6zJkzppivr68uWbKk6Xjz5s0a0H5+fqbYP//8owH91ltvmWJJSUm6VKlSetCgQabY0KFDdYUKFXRycrLWWuvk5GRdvnx5PXDgwHRfd0ruqf89JCUl6dKlS+siRYroiIgIU7xt27a6Tp06puNhw4bpYsWK6Zs3b5pit27d0q6urnrKlClWnysxMVFfvHhRA3rr1q1m71elSpV0QkKCKfbpp5/qUqVKZZh7drL4/SCENclJWu/sbPX3ePgMN12l1AltWPDBcKtbV+uoqOxNCdivH+OzPMcvWyilyqQ6bAWkzMRYDngrpRyUUhWBqsA+IAyoapxZ8TSGQZXLjW/GZiBlf+CuwDJbvIa8xNvbm9DQUOLj4wkNDbV6ySImJobdu3fTrl07EhMTTbcGDRpgb29vGh+htSYgIIAaNWrg6OiIvb09nTp1Ij4+ngsXzJdYffvtt82Oa9SowcWLFx8q75CQEG7dusXGjRvTvdTi6enJ6dOnWbBgAT4+Pty8eRNfX188PT0tLpP4+/sTFhZmduvZs2e6OYSFhQHg5eVlirm4uFidDZJZDRs2NPvLukqVKlSoUIGKFSuaxcLDw80uBwE0atTIrA8YXn8KOzs7KlWqxKVLl0wxHx8fzp8/z5YtWwDYvHkz58+f56OPrI/6Tu/57OzsqFixIrVr18bV1dUsj9TPt3HjRpo0aYKrq6vp31GhQoWoXbs2+/fvN/Vbs2YN9evXp3Dhwjz11FOUM85HO3HihFkOb731ltlZnho1anD9+nWL90aIXOXgYDj3s0U49l5BWkxaxqlr92fC1agBq1aBi4tF91zF1lM1FwC7gepKqYtKqe7AOKXUIaXU38BbwGcAWut/gEXAEWAt8LHWOkkbxjT0w3Am4iiwyNgX4Evgc+PgSjfgRxu+vDzBy8uLu3fvMnToUKKjo3nfyj6ut2/fJikpib59+2Jvb2+6OTg4kJCQwL//GoacBAQEMHDgQFq1asWyZcvYt28fU6dOBbC47l2kSBGz46efftqiT0a8vb35448/+Pbbb3F3d89woGOhQoXw9vZm9uzZnDlzhuHDh7Nz505WrFhh1q9KlSp4eHiY3cqWLZvu4169epVChQqZTrWnKFmyZDr3eDBr74u1mNba4gMydb+nn3463cdL/T5XqlSJhg0bMm/ePADmzZtHnTp1qFmzZpblmvr5bty4wcKFC83+Hdnb27N582bTv6OwsDC8vLwoV64cP//8M7t37zZt/Z2Zf0fW3hshco1j/nB0vEU4KdkO7ykh7Dpxf8bXs8/CunXg5mbLBB+NrWdbdLASTvcDXms9GrAYkaa1Xg2sthI/w/0ZGcIKZ2dnmjdvjr+/P23btsXZ2dmiT5EiRUzrPli7Dp7yAbt48WLatm1rNmjwyJEj2ZJ3xYoVqVOnDv7+/hbjMzKilMLX15evv/6aY8eO0aJFi0fOoXTp0kRFRREbG2tWQFy/ft2sX8GCBS0+zFIPeMxp//3vf+nRowdjxozh119/ZeLEidn2XMWKFcPLy4vhw4dbtKWMR1m6dCkmarQzAAAgAElEQVQlSpRg4cKFpoGt58+fz7achLCZc8Hwx+dWm3rPncHyA/d/H5UoARs25J5FoB4kNwyYzFsecpBibtSnTx/i4+Pp3bu31XZnZ2fq1avH8ePHGTFiRLqPExsbi4ODg1ksKCgoS3NNbeDAgQQHB9OlSxer7QkJCURHR1v8dXry5EkAs3UiHsUrr7wCwPLly2nfvj0Ad+/eZcOGDWan7suVK0dUVBSXLl3C3d0w4Wf9est163NK69at+fjjj/H29iY5OTndS0BZoVGjRixatIiaNWtanLFJERsbi729vdmMmOz8dySETVxZD3u6WW0aEfo/5mzuYTouVMiwemS1ala750pSPORDDRs2pGHDhhn2GTduHI0aNcLOzo42bdpQqFAhLly4wKpVqxg9ejTVqlWjSZMmTJ48mbp161K5cmWCgoI4depUtuXdrl27DGdnREREUK1aNbp27cpbb71F4cKFOX78OGPGjMHd3Z1WrVqZ9T9+/LjFlMqCBQtSq1Ytq49fs2ZNvLy86NOnD5GRkZQpU4bx48fj5ORk1q9p06Y4Ojri4+PDwIEDOXv2rMVMiZxUsGBBOnXqxNSpU+nQoYNFsZWVPv/8c3755Rc8PT3p378/7u7uXLt2ja1bt9KgQQM6dOhAkyZNCAgIYMCAAbz//vvs2rWLX375JdtyEiLb3dgH21sbZlikMX1jb75eev9MnIMDLF8OtWvbMsHHJ8WDsKpBgwZs27aNkSNH0rlzZ5KSkihfvjxNmzY1/QU/YsQIwsPDGWaciNy6dWsmT55sdRyFLbi6ujJo0CBWr15NcHAwkZGRuLu788477zBs2DCL1SO/+OILi8eoXLlyhgVQYGAgffr0YcCAAbi4uPDxxx/zyiuvmKavAhQvXpwlS5bwxRdf0LJlS2rXrk1wcDA1atTIuhf7mFq2bMnUqVPx8fHJ1ucpXrw4e/bsYejQoXz22WfcuXOHMmXK0KBBA1588UUAmjVrxnfffccPP/zA7NmzefXVV1m5ciXV8tKfYUKkuHMYtrxrWNMhjSX7WtMvcAopKwsoBQsWGLbYzmuUtrLyXn7j4eGhU4/8Tu3o0aM899xzNs5IiOw1aNAgFi5caFogSjwa+f0gzNw9AxsaQOwVi6atR9/gne/WEZ9wfyuACRNg4EBbJnifUuqA1trjwT2tkzMPQuQjx48f58iRI0yfPp2RI0dK4SBEVom9Ar83sVo4/H3hBVpMWmZWOPToAZ9bH0uZJ0jxIEQ+0qtXL/bu3YuXlxeffPJJTqcjxJMh/hb8/rbVZadPXq3C22PXExFzf2yRpydMnZqLV4/MBCkehMhHUhaHEkJkkYS7ho2uIg5bNF285U6TMRu4FnF/v51q1SA0FOztbZlk1pNzlkIIIcSjSIozbK19c69F040oN5qM2cD5GxVMsWLFDKtHFi1q0T3PkeIhE2RQqRAiLfm9kM8lJ8GuD+HaJoumyNhCNP1uLccu3x9Ma28Pv/4KxpXk8zwpHh7A3t6e2NjYnE5DCJHLpCxuJfIhrWH/x/DvEoumuHsOeE1czoGz5hMZZs2CNBsB52lSPDxAyZIluXTpEjExMfKXhhACrTUxMTFcunTpsfY1EXnYIT84NdMinJD4FG0mh7L1aEOz+P/+B9262SQzm5EBkw+Qsuzw5cuXSUiwXC1MCJH/2NvbU6pUKbNlyUU+cWIaHB5ltanbzEBW/dncLDZoEFjZ2iXPk+IhE1xdXeWXhBBC5HfnF8H+flabBvzsT/CuTmaxfv1g7Ni8PSUzPXLZQgghhHiQq5tg94eA5eXrMcsH8/3aAWYxHx/4/vsns3AAKR6EEEKIjN06YJiSaWWjqx+3+DBk4bdmsQ4dDAMkn+QFXJ/glyaEEEI8pogjsLkpJN61aFp+4H16/TiTlI2uAFq0gPnzoUABG+aYA6R4EEIIIayJPAGbGkH8DYumHcdfw3tKCEnJ94cOvvMOLFyY91ePzAwpHoQQQoi07p6F3xtB3FWLpkP/Ps/7E1YQe8/JFHvtNcMiUA4Otkwy50jxIIQQQqQW/S9s8oSYixZNZ65XpOl3a7kTc3+N6ZdegpUrwcnJovsTS4oHIYQQIkXsFUPhEH3Oounfm+XwHP07l2+7m2JVqsC6dVCkiEX3J5oUD0IIIQRA3HXDGIe7pyyartwujefo3802unJ3hw0boFQpG+aYS0jxIIQQQsTfgt/fhsijFk3hkcVpNGYTp65VNcXc3GD9eqhQwYY55iJSPAghhMjfEiJhy7tw5y+Lplt3i9J4zEaOXqphirm4wJo1UKOGRfd8Q4oHIYQQ+VdiDGxpDjf3WTRFxLjy9tj1/H3hJVPMwQGWLYNXXrFlkrmPFA9CCCHyp6R42NYKwrdbNN2Nc+bdcWvMtta2t4clS8DT05ZJ5k5SPAghhMh/khNgZ3u4ut6iKfZeQd6fuILdJ+ubYnZ2EBwM771nyyRzLykehBBC5C/JSbC7K1xcZtF0L9Ge1gG/suXIW6aYUhAYCG3a2DDHXE6KByGEEPmHToawXnB+gUVTYlIBvH8IYe1f75rFZ8yAzp1tlWDeYNPiQSk1Vyl1XSl1OFVsvFLqmFLqb6XUUqVUEWO8glIqVil10Hibkeo+tZVSh5RSp5RSk5UybHqqlCqmlNqglDpp/FrUMgshhBD5ktZwYACc/tGiKTlZ0W1mIEv3tzaL+/tDz562SjDvsPWZh0CgaZrYBuB5rfWLwAngq1Rtp7XWtYy33qni04GeQFXjLeUxBwObtNZVgU3GYyGEEPmd1vCnL5z4wWpz77kzCNr5oVnsm29gwABbJJf32LR40FpvA26lia3XWicaD/cA5TJ6DKVUGcBVa71ba62Bn4CWxuYWwHzj9/NTxYUQQuRXWsPfw+DYRKvNn/08idmbzU8vDB8OQ4faIrm8KbeNefAB1qQ6rqiU+lMptVUp9box5g6k3q3kojEGUEprfQXA+LVkek+klOqplNqvlNofHh6eda9ACCFE7nL4a/jnW6tNwxePImDtZ2ax776DUaNskVje9dSDu9iGUmookAgEGUNXgGe11jeVUrWB35RSNQFl5e76YZ9Paz0LmAXg4eHx0PcXQgiRBxz5Dg6NtNo0+rchfPPbMNOxUjBtGvTubbW7SCVXFA9Kqa5Ac6CR8VIEWut4IN74/QGl1GmgGoYzDakvbZQDLhu/v6aUKqO1vmK8vHHdVq9BCCFELnPMHw5aH/o2YdVAhi3+hpS/RwsUgPnzoVMnG+aXh+X4ZQulVFPgS8BLax2TKl5CKVXA+H0lDAMjzxgvR0QppeoZZ1l0AVIm6y4Huhq/75oqLoQQIj85MQ3++Nxq0+R1/fENHk9K4fD004aVI6VwyDybnnlQSi0AGgLFlVIXgZEYZlc4ABuMMy73GGdWvAGMUkolAklAb611ymDLPhhmbjhiGCORMk5iLLBIKdUduAC0tcHLEkIIkZsc80+3cJi5qSef/vQ9KYWDk5Nhr4rGjW2Y3xNAGa8S5GseHh56//79OZ2GEEKIx6E1HP4GDo2w2jxvaze6z/4RrQ0n3QsVgrVroX59q92faEqpA1prjwf3tC5XjHkQQgghHovWcPBLODreavMvOzrx39lzzAqH9euhXj1bJvnkkOJBCCFE3qaTYX8/ODndanPwrg50mxlIsi4ASOGQFXJ8wKQQQgjxyJITYc9H6RYOszf/l87TfiYp2fC3squrFA5ZQc48CCGEyJuS7sGujvDvEqvNAWs+5bNf/EkZHOnqCuvWSeGQFeTMgxBCiLwnOTHDwuHrpcOkcMhGcuZBCCFE3qKTDZcq0ikcvlwwlnErvzQdS+GQ9aR4EEIIkXdoDft6w7lfrDZ/HDiFaRs+Nh27uRkKh9q1bZVg/iDFgxBCiLxBa/jjMzg922pz91lzmLu1u+m4TBnYsAFq1rRVgvmHFA9CCCHyhr+HwfHvrTb1nTfVrHCoUAE2boTKlW2UWz4jAyaFEELkfodHp7ut9sCgCUzf2Nd0XL06bN8uhUN2kuJBCCFE7nZkvOGsgxUjQv/HpNUDTccvvQTbtkG5cla7iywixYMQQojcSWv4ewQcHGS1eezyL/l66XDTcd26sHkzlCxpqwTzLxnzIIQQIvfRyXDgMzgx2Wrz5HX9+WrhGFLWcahaFVatgqJFbZhjPibFgxBCiNwlORH2/hfOzrfaPGdzdwb8HEBK4VCsmKFwcHOzYY75nBQPQgghco+keOPKkb9abZ65qSd9500z7Y5pbw9LlxrOPAjbkTEPQgghcofEaNj6frqFw7iVvvSeO8O0OybA3Lnwxhu2SlCkkDMPQgghct69O7DlPbixy2rzVwu/ZezywaRcqgAYMQI+/NBG+Qkz6Z55UEqtV0pVT3WslFIjlFKl0/R7SSl1IjuTFEII8QSLuw6b3kq3cOg7bypjl39F6sKhQwfw87NNesJSRpctGgOF0/QdCZRN068gIEtxCCGEeHjR/8LGN+D2QYumxKQCfDjtZ7MFoABee81wuUIpi7sIG3nYyxbyoxJCCJE1Ik/C740h5oJFU3zC07T7YRHLD7Qwi3t4wG+/QcGCtkpSWCMDJoUQQtjenUOw8XWrhcPdOGeajV9tUTi0bAlbtkDx4jbKUaRLigchhBC2dWMvbHwT4q5ZNN2OLkLjMRv5/Z9GZvGBAyE0FJydbZWkyMiDLlvUV0ql1Hh2gAZeSzNo8j/ZkpkQQognz9VNsK2FYVpmGtciSvL22PX8feElU8zODqZMgT59bJmkeJAHFQ+TrMSs7YeqsyAXIYQQT7ILSwwLQCXfs2y68QyNx2zk5NVqppiLCyxaBO++a8skRWZkVDxUtFkWQgghnmynZkNYb8OeFWmcuFKVxmM28u/NZ02xcuVg5UrDLpki90m3eNBan7dlIkIIIZ5AWsOR7+Cvr6w2/3X+Rd4eu57rkaVMsbp1DUtOlyljqyTFw3qkFSaVUu9jGOtwDViqtY7K0qyEEELkfToZ/vSFY9augMPOE/VpPn4ld2Lub4XZpQvMnClTMXO7jFaY/FIptT1NzF4ptQ34DfgOCAQOK6XSLhyVLqXUXKXUdaXU4VSxYkqpDUqpk8avRY1xpZSarJQ6pZT6Wyn1cqr7dDX2P6mU6poqXlspdch4n8lKyTIiQghhc8mJsMcn3cJh9cF3aTJmg6lwsLODCRMgMFAKh7wgo6marYCdaWKfAA2AbwBXwANIAoY+xHMGAk3TxAYDm7TWVYFNxmOAd4GqxltPYDoYig0Mq13WBeoAI1MKDmOfnqnul/a5hBBCZKfEGNjeOt0ttX/Z0YkWk5YRe88JAFdXw/iGgQNl1ci8IqPioTKwJ03MGzirtR6ptb6rtf4DGAs0yewTaq23AbfShFsAKf/K5gMtU8V/0gZ7gCJKqTLAO8AGrfUtrfVtYAPQ1NjmqrXerbXWwE+pHksIIUR2u3cbNr8Nl1ZYbZ68rj9dZvxEYpI9YNhKe+9emVGR12RUPDgBd1IOlFIuwP8BG9P0Owa4P2YepbTWVwCMX0sa4+7Av6n6XTTGMopftBK3oJTqqZTar5TaHx4e/pjpCyGEIOYSbHgdwtOetDYYvngUn/70PVobPnrq1oU9e+A/slpQnpNR8XAGwyWBFE0w7G2xKU2/wkBkFueVwtoJLP0Iccug1rO01h5aa48SJUo8RopCCCGIOAbr60PEPxZNycmKvvOm8s1vw0n5Nd2wIWzYAMWK2TZNkTUymm0xD/BTSiVimFXxPyAcWJWm31vA8cfM45pSqozW+orx0sN1Y/wi8EyqfuWAy8Z4wzTxLcZ4OSv9hRBCZJcbe2HrexB/06LpXqI9H077hcV725lizZoZlpp2dLRlkiIrZXTmYTKwABgD/AzYAx201qY1RZVShYGuwOrHzGO58XEwfl2WKt7FOOuiHhBhvKyxDnhbKVXUOFDybWCdsS1KKVXPOMuiS6rHEkIIkdUur4VNnlYLh6hYF94dt8ascGjb1rCGgxQOeVtGi0QlAr2UUgMAZ631DSvdooFqPMRlC6XUAgxnDYorpS5imDUxFliklOoOXADaGruvBpoBp4AY4CNjbreUUl8DYcZ+o7TWKYMw+2CY0eEIrDHehBBCZLWzQbCnG+hEi6ZrESV5d9wa/jxnmmFPt24wezY89UgrDIncRBkmJeRvHh4eev/+/TmdhhBC5B3H/OGPz602nb5WiXe+W8fpa1VMsX794PvvDes5iJynlDqgtfZ41PunW/8ppbo8zANprX961CSEEELkEVrDwcFwdJzV5oPnX6Lpd2u5FnF/8+Uvv4QxY2QNhydJRiePArk/U+FBP/KUNRWEEEI8qZITYV8POBNotXnzkYa0nPQbkbGFTbH//Q+GD5fC4UmTUfEQjaEoWAaEANsAy+3QhBBCPPkSY2BHe7i80mpz6L4P+HDaL8Qn3F9betw48PW1VYLCljK6+lQC8MEw8HAxcBQYBdTQWkenvdkgVyGEEDkh/hb83iTdwmH6xt60n7zQrHD44QcpHJ5k6RYPWus4rXWo1roNhhUfh2DYSXOHUuq0Umq0UkrWBRNCiCfZ3TOwoT7c2GW12W/JSPrOm0ayLgAYLk/MmWMYICmeXJmaMKO1vothTMNPxnUVhgCDgOeA1tmXnhBCiBxzYw9s9YJ4yyX8k5MVHwdOZcamPqZYgQLw00/QsaMtkxQ5IdOzbZVS5YD2GDbHqg3sw7CIlBBCiCfNhVDY3RmS4iya4hOeptO0IJbsa2OKFSwICxZAS9mKMF/IsHhQSpXAsGBTB6A+cAjD4Mm2Wutz2Z6dEEII29Iajk6Ag4OsNkfEuNLKfymbj3iaYsWKwYoVUL++rZIUOS2jdR7WYdi34jSwEPiv1vpx97AQQgiRWyUnwv7+cGqG1ebzN57lvfGr+Ofi86ZYxYqwZg1Ur26rJEVukNGZhybAXeOtGdBMZTBRV2tdJ91GIYQQudu9CNjpDVfWWm3ef6Y2709cwdU7ZUyx2rVh5UooXdrqXcQTLKPi4SfS2c5aCCHEEyTqFGx9HyKPWW1edsCLjlODiYl3NsWaNoXFi8HFxVZJitwko42xutkwDyGEEDnh2mbY3gbu3bLaHLDmUwYGTTRNxQT46COYORPs7W2VpMhtZIsSIYTIr07OhN/ftlo4JCXb8clP3/PZLwFmhcOwYfDjj1I45HeyMaoQQuQ3yYmGHTFP/GC1OSrWhQ5TF7Dqz+ammFIwZQr07WurJEVuJsWDEELkJ/fuwI52cHWD1eaz1yvw/sQVZjMqHBwgOBhay5KAwkiKByGEyC/unoEtzSHyqNXmbcde54OAJdyIKmGKFSkCy5fD66/bKkmRF0jxIIQQ+UH4LtjW0upS0wBzNnen77xpJCQ9bYq5u8PatfD881bvIvIxKR6EEOJJdy4E9nSD5HiLpqRkO74ImkDA2gHA/bV8atY0LP70zDO2S1PkHQ+zt0UbDJtglQMKpm2XRaKEECKX0RoOfwOHRlhtjowthPcPIaz5q5lZvF07w4wKWcNBpCdTxYNSyg8YAfwFHAHuZWNOQgghHldSPOztAed+ttp8/sazNJ+wksP/vmCKFSgA48fDgAGG2RVCpCezZx66A2O11kOyMxkhhBBZIO46bP8AwndYbd53+hW8Ji7nWsT9daVLlYJFi+CNN2yVpMjLMls8FAI2ZWciQgghssCdQ4alpqPPW21evLcNXWfMJ/aekylWv75hqemyZW2VpMjrMrvCZAjQNDsTEUII8ZguLof19dMtHMYsH0z7HxaaFQ79+8PmzVI4iIeT2TMPm4DvlFLFgQ3AnbQdtNarszIxIYQQmaQ1HPkO/hqCtf0MExKfotfcmczb6mOKubjAnDnQvr0N8xRPjMwWDwuNXysAXa20a6CAlbgQQojslBRnHBj5i9XmG1FufBCwhG3H3jTFataE0FD4z39slaR40mS2eKiYrVkIIYR4eNEXDEtN39xrtfmfizV4f8IKzoZXMsU6dTLsiOnsbPUuQmRKpooHrbX1C2hCCCFyxuU1sOvDdLfSXvnne3ScGkxUrCsATz8NkydDz54yDVM8vodaYVIp9RTwLNYXiTqSVUkJIYRIR3IiHPKDf0an22X8yi8YHDLWtJV2+fKwZAnUrm2jHMUTL7OLRNkDkzGMd3BIp9sjj3lQSlXn/rgKgEoYFqUqAvQAUhZjH5IyMFMp9RWG9SeSgE+01uuM8abA98Z85mitxz5qXkIIkavEXoWdHeD6FqvN9xLt6fnjLOZv62aKNW4MISHg5mabFEX+kNkzDyOA5hg+rIOAj4Fo4EOgMtD/cZLQWh8HagEopQoAl4ClwEeAv9Z6Qur+SqkagDdQEygLbFRKVTM2TwWaABeBMKXUcjkrIoTI865tMRQOcVetNl+85U67yYvYfbK+KTZoEIweDU/JLkYii2V2nYd2gB+wyHi8T2v9k9b6bWAH0CILc2oEnH7AOIsWQIjWOl5rfRY4BdQx3k5prc9ore9hWJ8iK3MTQgjb0hqOToDfG6VbOKz96x3+b8ifpsLByQkWLoTvvpPCQWSPzBYPzwAntNZJQBxQNFVbEPBBFubkDSxIddxPKfW3UmquUirled2Bf1P1uWiMpRe3oJTqqZTar5TaHx5ufYtaIYTIUYkxsKsT/OkLOtmiOSnZjmGLv6bZ+NXciCoBQOXKsGePYXMrIbJLZouHKxjGHwCcBVKvfl45q5JRSj0NeAGLjaHpxsevZcxhYkpXK3fXGcQtg1rP0lp7aK09SpQo8Vh5CyFElrt7Dja8BucXWG2+FlGSJmM2MPq3YWht+FXetCmEhcELL1i9ixBZJrMntLYArwMrgNnABKVUFSAeaI/5mYLH8S7wh9b6GkDKVwCl1GxgpfHwIoazISnKAZeN36cXF0KIvOHq77CzHcTftNq89egbdJiygCt37q8pPXgwfPONYWdMIbJbZouHoUBxAK11gFJKAW0AR+AHYFQW5dOBVIWIUqqM1vqK8bAVcNj4/XIgWCk1CcOAyarAPgxnHqoqpSpiGHTpDXTMotyEECJ7aQ3Hv4c/vwCdZLXLpNWfMWjBOJKSDb++HR1h7lzw9rZloiK/y+wiUVeBq6mO/QH/rExEKeWEYZZEr1ThcUqpWhguPZxLadNa/6OUWgQcARKBj43jMVBK9QPWYZiqOVdr/U9W5imEENkiMRr29U53menYewXpMWc2QTs/NMWefRZ++w3+7/9slaQQBkprq0MCrHc2TJGsjeHSwFyt9VXj5YtrWuuobMox23l4eOj9+/fndBpCiPwq4ijsaAMR1meVX7jxDC39f+PPcy+bYm+8YdifQoZsiUehlDqgtfZ41PtndpEoF2AuhlkVicb7rcVwNuJb4ALwxaMmIYQQ+da5BbCvh+HMgxVbjrxJux8WER5Z0hTr2xcCAsDe3lZJCmEus7MtJgH1gcZAIcxnNawGmmZxXkII8WRLioewvrCrY7qFw/drP6HJ2A2mwsHeHmbPhqlTpXAQOSuzAyZbA59qrTcbV4BM7TxQPmvTEkKIJ9jdc7CjLdyyfrk0Os6J3vNm8MuOzqZY6dKG/Snq17d6FyFsKrPFgyNgfc6Q4UyE9WHBQgghzP37G+z5CBLuWG0+drk6HwQs4cilmqZYnTrw66/gbnXJOyFsL7OXLcKALum0tQF2ZU06QgjxhEqKg/39YXurdAuHkN3teWV4mFnh0K0bbN0qhYPIXTJ75mEYhs2nNmJY/VEDzZRSn2EoHt7I6M5CCJGvRZ6Ene3h9p9Wm+8l2vPZL/5M29CXlCFlBQrApEnQvz8oa2vnCpGDMrvOww6lVCNgLDAFw7/u/wF7gMZa67DsS1EIIfKws0EQ1hsS71ptPn/jWdp+v5iwM3VMsVKlDBtbvfmmrZIU4uFker81rfVO4HWllCOGjbHuaK1jsi0zIYTIyxKjYf8ncGZuul2WHfDCZ9Zcbt11M8Veew0WLYKyZdO9mxA57qE3a9VaxwKx2ZCLEEI8GW79YZiCGXncavO9RHt8g8czed0npJ75PmAAjBsn0zBF7pdu8aCUGvEQj6O11l9nQT5CCJF36WQ4OgH+HgbJCVa7nLpamfY/LOSPc7VNMWdnw/4Uso22yCsyOvPgh+EMQzTWt7pOTQNSPAgh8q+YS7C7K1zblG6X4F0d6D13BlGxrqbYf/5jmIb53HO2SFKIrJFR8XAGeBY4AIQAS7XWkTbJSggh8pJ/l8Le/8K9W1abY+Id6T//B+Zu9SH132I+PvD99+DiYqM8hcgi6a7zoLWugmFJ6n8wnFW4qpT6VSnV1jhoUggh8rfEGNjXC7a3Trdw+PNcLWoPO8Dcrd1JKRyKFDEMivzxRykcRN6U4SJRWuv9WusvtNbPYti/4iqGqZrXlVJBSilZ30EIkT/d/hvWesCpWel2mbBqIPVG7uHY5fvXJF5/Hf76C9q2tUWSQmSPh5mquQ3YppQaAIwGPsOwbPW2bMpNCCFyH63h5DT4YyAkx1vtcvl2GbrOmM/Gw01MsQIFwM8PvvrK8L0QeVmmiwel1GuAN4YVJQsBocD0bMpLCCFyn/ibsLc7XFyWbpdlB7zoPutHbt4tbopVrAjBwVCvni2SFCL7ZVg8KKVexlAwtAdKAWsxnHFYLgtECSHylWtbYfeHEHPRanPsvYJ8/sskZmzqTepBkZ07w5Qp4Opq9W5C5EkZrfNwHKgI/A6MBH6V2RZCiHwnMQb+GgbHAzDMSrd0+N+aeE8J4Z+Lz5tihQvD9OnQoYON8hTChjI681AViANqAy8D41QGu7NorUtmbWpCCJHDru+AvT4QdTLdLtM39mZg0ERi7zmZYg0awC+/QPnytkhSCNvLqHj4n82yEEKI3CQxGv4aCscnk97ZhtvRReg+60eW7m9tismgSJFfpFs8aK2leBBC5D/Xt8EeH7h7Ot0u2481oNO0IP69+eCaO8QAACAASURBVKwpJoMiRX7y0BtjCSHEEykxFv4aYhzbYF3cPQdGLBnFpNWfk5R8/9dnhw6G8Q2FC9siUSFynhQPQghx6wDs6gyRR9PtsudUXT6aOc9swSdnZ5g6Fbp0gQyGhAnxxJHiQQiRfyUnwj9j4PAo0IlWu8TeK8jwxV/jv+YzkvX9gQwvvwwLFkC1arZKVojcQ4oHIUT+FHkCdneBm3vT7bLzRH18Zs3lxJXqZvHPP4dvvwUHh+xOUojcSYoHIUT+opPhxDQ4+CUkWV/rLu6eA0MXjyZgzQCzsw0vvAAzZ8Krr9oqWSFyJykehBD5x53DsK8n3Nidbpc/z9Wi8/SfzRZ8cnIyTMEcMADs7W2QpxC5XIa7atqaUuqcUuqQUuqgUmq/MVZMKbVBKXXS+LWoMa6UUpOVUqeUUn8bl9JOeZyuxv4nlVJdc+r1CCFyif9v777jq67uP46/PhkECFMZshHcrVYEUVGrWAdW666ltT+rtUXrKK7ibGvVqrUuVBwUrNpasbUOtEqL4kSUpYjsISOEEUaABDLv5/fH+SaGmMGNyc16Px+PPO69557v935Ovsm9n3u+53tO0c4wb8ObAypNHIpjSfzxlZs54nef7JI4nH46zJ8Pv/mNEgeREg2x52Gou28s8/hG4G13v8fMbowe3wCcSpgFc1/gCMIiXUeY2R6E6bQHEWZ3mWVmE919SyIbISINxLopMP1SyFlaaZWl6/pz4RPPMm3JkNKyHj3gkUfgrLN0JYVIeQ2q56ESZwLPRPefAc4qU/6sBx8DHcysG3AKMNndN0cJw2RgWKKDFpF6lr8Jpl0EU75XZeLwxNuXcujNn5UmDmZwxRWht+Hss5U4iFSkofU8OPA/M3PgSXcfC3R197UA7r7WzErW0OgBrC6zbUZUVln5LsxsBDACoHfv3uWfFpHGyh1WToBZIyE/q9JqCzP359LxT/L+wuNKyw48EMaNgyFDKt1MRGh4ycPR7p4ZJQiTzWxhFXUr+j7gVZTvWhASk7EAgwYNqnjyehFpXHJXwYxfQeYblVYpKErlrldv5u6JN1FQFK61TE2FW26BG2/U5Zciu6NBJQ/unhndbjCzl4HBwHoz6xb1OnQDNkTVM4BeZTbvCWRG5ceXK3+3jkMXkfoUK4Ylj8Gcm8KiVpX4YOExjBg/dpdZIocMgb/8BQ46KBGBijQNDWbMg5mlm1nbkvvAycAXwESg5IqJnwGvRvcnAhdGV10cCWyNTm/8FzjZzDpGV2acHJWJSFOUPRcmHwOzfl1p4pCd255fjhvLcXe+V5o4dOwIY8fCBx8ocRCJV0PqeegKvGxhdFIK8A93n2RmM4B/mtklwCrgh1H9N4DvA0uBHcDFAO6+2czuAGZE9W53982Ja4aIJERRLsy9HRY+UOnU0gAvTj+Xq555hHXZ3UrLLrwQ/vxn6NKl0s1EpArmrtP9gwYN8pkzZ9Z3GCKyu9a8DjOvhNyVlVfZ3J0rnh7Dq7POKi3bf/+w+uXQoYkIUqThMrNZ7j6opts3pJ4HEZGq7cgIV1GsfqnKao+/dRk3TriHbTvDGtlpaXDrrWGiJw2IFPnmlDyISMMXK4TFj8Lnv4OinEqrLcrcj1+MG8eHi44tLTvtNHj4YejXLxGBijQPSh5EpGFb93YYDLl1fqVV8grSuGvizfzptRtKL7/s0wdGj4YzztBETyK1TcmDiDRMuSth9nWw+t9VVps890Qu/+tjLF2/LwAtWoTTEzffHBa0EpHap+RBRBqWop2w4F6Yfw8U51VabV12V675+4NMmDackrnhTjwRxoyB/fZLUKwizZSSBxFpGNwh4xWYfS3krqi0WixmPDnlUm564W627ugAwF57wUMPwfnn6xSFSCIoeRCR+rd1QRjXsO6tKqt9tPgornrmEWavGAhAUlJYxOqOO6B9+0QEKiKg5EFE6lPBVpj7B1j8SJUTPa3dshejJtzLc1MvwD1MjDtoEDzxBAwcmKhgRaSEkgcRSTyPwfKnw1oUeRsqrVZYlMJDk67mjld+y/ad7QBo1w7uugsuuwySkxMUr4jsQsmDiCTW+nfg09/A5llVVps05xSu/ttDLFp7QGnZj34EDz4I3bpVsaGI1DklDyKSGFvnw6ejIPM/VVZbtr4f1/z9QV6b/QNKrqLo1w8eewxOOSUBcYpItZQ8iEjd2rkO5v4elo0LpysqkZvXmj++egsPvHkt+YUtAUhNhVGj4JZboFWrRAUsItVR8iAidaMwBxbeDwv+XOlS2SWe/2g4o56/l4zNvUrLjj8+9DYceGAdxykicVPyICK1K1YIy8bD3Nsgb32VVWcuH8h1z93P+wuPKy3r2xfuuw/OOUdzNog0VEoeRKR2uEPGy/DZTbB9cZVVV27szU0v3M2EacNLL71MTw9TSl97LbRsmYiARaSmlDyIyDe34UP4bBRsnFZltezc9vzx1Vt45H9XlY5rALjwQrj7bujeva4DFZHaoORBRGpuyxyYcytkvl5ltYKiVMZMvoI7X7mVzTl7lpZ/97tw771wxBF1HaiI1CYlDyISv21LwhUUK5+vturzHw3n1n/dyfIN/UvLvvOd0NMwbJjGNYg0RkoeRGT37ciAubfD8qfAi6usOmXeUEY9fy+zvhxUWtavX1iHYvjwsC6FiDROSh5EpHo718P8P8GSxyCWX2XVz1cdzA0T/sSkOcMomeSpa1f47W/hl7+EFi0SEK+I1CklDyJSubysME/D4keheGeVVVdk9eG2l27jbx/8HzEPi0506gQ33ACXXw6tWyciYBFJBCUPIvJ1+Ztgwf2w+OFqJ3hav7ULd75yK2OnjKCgKA2Ajh3hN7+Bq66CNm0SEbCIJJKSBxH5Sv4mWPgQLBoNRdurrJqd2557Xx/F6P+OZEd+OgDt28N118HIkWH1SxFpmpQ8iAjsyAxTSS99stqehty81jz8v19z72ujyN7REYC2beHqq+Gaa0Kvg4g0bUoeRJqznC/DQMjlf4VYQZVVdxa0ZMzkK7j39VFkbesChFkhR44MvQ177JGIgEWkIVDyINIcbZ0P8+6Blf+o9pLLvII0nnj7Mv70+g2sy+4GhMGPV10F118fBkWKSPOi5EGkOdk0A+bdBRmvVFs1v7AFY6eM4J7XbiRzSw8g9DRcfnnoaejata6DFZGGSsmDSFPnDhveDUnDureqrb4jvxVPTrmU+9+4jjWbewJhTMNVV4UxDeppEJEGkTyYWS/gWWAvIAaMdffRZnYb8EsgK6p6s7u/EW1zE3AJUAz82t3/G5UPA0YDycA4d78nkW0RaTBixbDmtTCmYdPH1VbPzm3Po5OvZPSkkWzc3hmADh3CmIaRIzUQUkS+0iCSB6AIuM7dZ5tZW2CWmU2OnnvQ3e8rW9nMDgKGA98CugNvmdl+0dNjgJOADGCGmU109/kJaYVIQ1C4HZY9FeZoyFlebfUNWzvz4KRreGzy5Wzb2R6Abt1CT8Pll4fLL0VEymoQyYO7rwXWRve3m9kCoEcVm5wJTHD3fOBLM1sKDI6eW+ruywHMbEJUV8mDNH05K2DRw7B8PBRuq7b6qo29uO+N6xn3zi/YWRCmfxw4MJya+OEPNY20iFSuQSQPZZlZX2AA8AlwNHClmV0IzCT0TmwhJBZl+2Ez+CrZWF2uvMLFfs1sBDACoHfv3rXXAJFEcoesqbDoIch4GTxW7SaLMvfjntdu5LmpF1BY3AIzOPvskDQcc4xWuRSR6jWo5MHM2gD/Bq52921m9jhwB+DR7f3AzylZbWdXDlS0Tp9X9FruPhYYCzBo0KAK64g0WMUFsOqFMBvkltm7tcnsLwdw18SbeXnG2cQ8mVat4NJLQtLQr18dxysiTUqDSR7MLJWQODzn7i8BuPv6Ms//BXg9epgB9CqzeU8gM7pfWblI45e3AZY8GVa3zFu3W5u88dmpPDTpaibPPQkw9twzjGe44gpdOSEiNdMgkgczM2A8sMDdHyhT3i0aDwFwNvBFdH8i8A8ze4AwYHJfYDqhR2JfM9sbWEMYVPmTxLRCpI64h/kZloyBlS9UuyQ2hMstn/3wQkZPGsnCzAMB6Ns3zM9w8cVhvgYRkZpqEMkDYWzD/wFzzeyzqOxm4Mdmdijh1MMK4FIAd59nZv8kDIQsAq5wD9PkmdmVwH8Jl2o+5e7zEtkQkVpTtBNWTghJw+ZZu7VJ5pZuPPq/K3lyyqVsztkTgGOPDb0M554LKQ3lP15EGjVz1+n+QYMG+cyZM+s7DJFg+1JY8gQsfwoKtuzWJtOXHc5Dk67mxU/Oo7C4Ba1bw09/GpKGQw6p43hFpNExs1nuPqim2+t7iEhDULQDVv8blo2HDe/t3ibFyfx7xrk89ObVfLz0SMDo3z8kDBddpEmdRKTuKHkQqS/u4XTEsvFhgardmJsBYHNOR8ZOGcGYyVeQsbkXycnhUsvLLoMTT4Skiq45EhGpRUoeRBItbyOseC6clsj+fLc3m7FsEGPeuoIXpv2IvMJW9OgBt90Gv/gF9KhqSjURkVqm5EEkEWLFsG5y6GVY8yrECndrs7yCNJ6f9mMee+tyZi4/HICTToJf/Qp+8AMNgBSR+qG3HpG6tG0xfPksLH8adq7Z7c0WrDmA8e9ewl/fv5jNOXvSqROMGgUjRkD//nUXrojI7lDyIFLbdq4Ll1iueA427/5VPDl56bzw8Y8Y/+4lTFtyFGAce2wYy3DuuZCWVnchi4jEQ8mDSG0o3A6rXw4Jw/q3dmuNiRLTlhzJuHd+wT8/OZ+cvLZ07QrXXx8mczrooDqMWUSkhpQ8iNRU0Q7I/E+Y9THzP1Cct9ubrt/ahWc/uJCn3vs5CzMPJDU1jGG4+GIYNkxjGUSkYdNblEg8ivMgc1JYlGrNa1CUu9ubFhUn85/PTuOp937OG599n6LiVA45BB4aBRdcoHUmRKTxUPIgUp2i3JAwrH4JMl/f7fkYSny64lCem3oBf5/6U9Zv3Yv27WHEpfDzn8Nhh2kJbBFpfJQ8iFSkIBvWvB4ShrWToHhnXJt/uaEv//joJzz30QUsWBMGLpxwAjxwSZjQqVWrughaRCQxlDyIlMhZEU5FrHkNNry723MxlMja1okXp5/H36f+lI8WDwGMI46A+6+F886D3r3rImgRkcRT8iDNl8dg0/SvEobsuXHvYktuB16acQ4Tpg3nnflDKY6lMHgw3HdfSBj69KmDuEVE6pmSB2le8jaGmR4z38TX/hfL3xD3LrbuaMers87khY9/xOS5J1EUa8GQIXDvn+Gcc6Bv39oPW0SkIVHyIE1brDgsPrX2zZAwbJqOEZahj2ecYta2Trwy6yxemnEOU+adQLGnMXQojH4EzjoLunWrm/BFRBoiJQ/StLhDzrLQu7DuLXzdFKwwu/TpeBKGjM09eGnGObw04xw+XHQMbdqmcMop8Jdr4bTTYM89az98EZHGQMmDNH471sCG92D9O/jaydiOlaVPxXsV5KwvD2Pi7DN4bfYP+HTFAA480DjtNPjD4zBkCKSm1m7oIiKNkZIHaXxyV8L692DDe8TWv0dS7rLSp+JNFnYWtGTKvBOYOPsMXv/0dHJjPfje9+DSG+Hkk2HvvWs3dBGRpkDJgzRsHoOt8yDrQ8iaSvG6D0nO+6pnIakGu1y8dl/enHMqb845lQ8XH8d3DmvFySfDi3fC4YdramgRkerobVIaloLssBLlxk8oXj8Vz/qIlNjW0qeTa7DL7Tvb8O6C45n0+TAmzRmGp/fnlFPg0ttgwlDo0KHWohcRaRaUPEj9Kc6DLXNg03QK10+naP0MWhUuKn26JokChDUkPll2BJPnnsRbX5zIoo1HcPSxqZx8FlzzGPTvrymhRUS+CSUPkhiF22DLHHzzbHZkfErxxtm0KZ5PkhUDkBr91EQsZny6cgDvLTiOdxccz+yM4xkwuB1Dh8Ijt8Ihh0ByTTMRERH5GiUPUrtiRbB9CZ49l9w1X5C3bi6pO76gfdJSIAxoTC+pW8Nv/0XFycxecRjvLTiO9xYex9LsYzh4YAeGDIHfjoQBAzRuQUSkLuktVmqmaAdsX0J+1gKyVy6kYNNC0vIW0DFlIalJBRjQJvqp0ajGMrbvbMO0pUcxdfHRfLT4GHJaHsGAw9tw9Gnw6F1hCmidhhARSRwlD1K5op14znK2ZiwlO2MZhZuXkrJzKW1Zwh4tV5JkThrQtaR+i9p52YWZ+zNj+eFMXzaYRZuPoX2fgzl8cArfGwmjBkJ6evX7EBGRuqPkoblyx/M3sXXtGrasWcWOrJUUbV1Fct4qWvlK2qeuolPrTAzoEP0AUMtLSa/e1JOZywcxfflgMnYcDnsOYr9vdeDQU+GGAdCzZ+2+noiIfHNKHoDtW/P5bGYunfZKp3NnSEur74hqpiDf2bIhm21ZWezYnEVedhaFOVn4zg1YwXrSijNJT8qkQ1ome6avJS2lYNfEoAW11ntQkSXr9mH2isOYm3EY25IHkNx5AHsf0JlDz4YbDtElkyIijUWTTB7MbBgwmnC13zh3v6eq+m2Lv+DQxW3YNqctK7P3IiunG9n53cgt7kIxbShOakMsqQ2enA6pbUhKTScptSWWkkZyahpJqWmkpKaR3KIlSampJCUnkZyURFJyyU8ylmTgMTxWXHrrJbdFBRQX5hMryidWmI8X5+NF+XjhDrwwFy/KwYpySIrlkOw5pPh2Wlg2rZK20Colm7ZpW2jXcgsdWmfTNbn4q9MIUOcJQUWytnVi7uqDWbD2YDYWfZvC1gfTuttB7P/tdhz2Azivn65+EBFpzJpc8mBmycAY4CQgA5hhZhPdfX5127ZrtZ12rbazH0tqHkBh9FMTRsI/6L+JFVl9WLj2ADK3H8B2O4DiNgeQ3u1Aeu/XlQNOgeN6K0kQEWmKmlzyAAwGlrr7cgAzmwCcCVSbPMiuYjEjY3NPvty4D1k79yHH+lPceh9a7NGfPfrsR78BrTl+b2jZsr4jFRGRRGqKyUMPYHWZxxnAEeUrmdkIYATAwGa6+FFOXjprt3ZnY24vsgv7sDOpN8Ut+5DSrjetO/dmj5696Nu/Jd/tpEshRUTkK00xeajoY86/VuA+FhgLcGifFl5Y7KQmF9V1bHUuJz+d7B2d2VbQmdyizuTRmcKkzsRadCEpvTtpHbqT3qk7Hbp3p3O3tuybDvvWd9AiItKoNMXkIQPoVeZxTyCzqg1SOh9C6k+nQ/5mYjvWsj1rHTlZaynYvomi/Fxi+Tl4YS4U5WDFYeBikueTRD7J5JNEAcmWT4rlk0QRZjHMikkihhEjyYoxixHzZGKejHtSuE8S7kkUeQuKY2kUeRrFnkYRacQ8jSJaEYsGa8aS20BKGywlHWvRhpTWHWiR3pG0th1o3aEj6R070rp9B9qkpIWJmUREROpIU0weZgD7mtnewBpgOPCTareyJGjZiaSWnWi/x8G037+OoxQREWmkmlzy4O5FZnYl8F/CpZpPufu8eg5LRESkyWhyyQOAu78BvFHfcYiIiDRF33DJIhEREWlulDyIiIhIXJQ8iIiISFyUPIiIiEhclDyIiIhIXJQ8iIiISFyUPIiIiEhczP1ryz40O2a2HVhU33HUo07AxvoOop4057aD2q/2N9/2N+e2A+zv7m1runGTnCSqBha5+6D6DqK+mNnM5tr+5tx2UPvV/ubb/ubcdgjt/ybb67SFiIiIxEXJg4iIiMRFyUMwtr4DqGfNuf3Nue2g9qv9zVdzbjt8w/ZrwKSIiIjERT0PIiIiEpdmnTyY2TAzW2RmS83sxvqOp66ZWS8ze8fMFpjZPDMbGZXvYWaTzWxJdNuxvmOtS2aWbGafmtnr0eO9zeyTqP0vmFmL+o6xrphZBzN70cwWRn8HRzWX429m10R/91+Y2fNm1rIpH3sze8rMNpjZF2XKKjzWFjwcvRd+bmaH1V/ktaOS9v85+tv/3MxeNrMOZZ67KWr/IjM7pX6irj0Vtb/Mc9ebmZtZp+hx3Me/2SYPZpYMjAFOBQ4CfmxmB9VvVHWuCLjO3Q8EjgSuiNp8I/C2u+8LvB09bspGAgvKPP4T8GDU/i3AJfUSVWKMBia5+wHAdwi/hyZ//M2sB/BrYJC7fxtIBobTtI/908CwcmWVHetTgX2jnxHA4wmKsS49zdfbPxn4trsfAiwGbgKI3geHA9+Ktnks+oxozJ7m6+3HzHoBJwGryhTHffybbfIADAaWuvtydy8AJgBn1nNMdcrd17r77Oj+dsIHRw9Cu5+Jqj0DnFU/EdY9M+sJnAaMix4bcALwYlSlybbfzNoB3wXGA7h7gbtn03yOfwrQysxSgNbAWprwsXf394HN5YorO9ZnAs968DHQwcy6JSbSulFR+939f+5eFD38GOgZ3T8TmODu+e7+JbCU8BnRaFVy/AEeBEYBZQc8xn38m3Py0ANYXeZxRlTWLJhZX2AA8AnQ1d3XQkgwgC71F1mde4jwjxOLHu8JZJd5Q2nKfwf9gCzgr9Fpm3Fmlk4zOP7uvga4j/Btay2wFZhF8zn2JSo71s3x/fDnwJvR/WbRfjM7A1jj7nPKPRV3+5tz8mAVlDWLS0/MrA3wb+Bqd99W3/EkipmdDmxw91lliyuo2lT/DlKAw4DH3X0AkEsTPEVRkejc/pnA3kB3IJ3QVVteUz321WlO/weY2S2E07jPlRRVUK1Jtd/MWgO3AL+r6OkKyqpsf3NOHjKAXmUe9wQy6ymWhDGzVELi8Jy7vxQVry/poopuN9RXfHXsaOAMM1tBOE11AqEnokPUlQ1N++8gA8hw90+ixy8SkonmcPxPBL509yx3LwReAobQfI59icqOdbN5PzSznwGnAxf4V3MVNIf29yckz3Oi98CewGwz24satL85Jw8zgH2j0dYtCINlJtZzTHUqOr8/Hljg7g+UeWoi8LPo/s+AVxMdWyK4+03u3tPd+xKO9xR3vwB4BzgvqtaU278OWG1m+0dF3wPm0zyO/yrgSDNrHf0flLS9WRz7Mio71hOBC6NR90cCW0tObzQlZjYMuAE4w913lHlqIjDczNLMbG/CwMHp9RFjXXH3ue7exd37Ru+BGcBh0ftC/Mff3ZvtD/B9wojbZcAt9R1PAtp7DKEr6nPgs+jn+4Tz/m8DS6LbPeo71gT8Lo4HXo/u9yO8USwF/gWk1Xd8ddjuQ4GZ0d/AK0DH5nL8gT8AC4EvgL8BaU352APPE8Z3FEYfFJdUdqwJ3dZjovfCuYSrUuq9DXXQ/qWEc/sl739PlKl/S9T+RcCp9R1/XbS/3PMrgE41Pf6aYVJERETi0pxPW4iIiEgNKHkQERGRuCh5EBERkbgoeRAREZG4KHkQERGRuCh5EKkH0fXUX0Yr2+1Tg+27mNlt0TTjdSLa/8YabNcmatdFcW73tJnNjPf1vgkze83Mfl8L++kbtfn0OLcbZGabzKz9N41BJJGUPIjUj6OAvtH94TXYvgvw+zL7kDiZ2RHAUOCRWtjdWsIx/TCejdx9JmG+gWtqIQaRhFHyIFI/fkxYW+KT6L4k3q+BV929opUH4+JhNcaPPaxSGq+/ApeVmSZbpMFT8iCSYGaWDPyQMCXsU8BBZnZIBfX6mNnzZrbRzHaY2edm9pPoVMXcqNo7UXe5R9tcFD1uU25fK8zsvjKPTzOzyWa2wcy2mdnHZnZyDdtzrpktNrOdZvY+cEAl9X5hZvPMLN/MVprZqGr2283MnjKz5dG+F5vZndF08iV1ZpjZXyvY9hkzm13FvtsCZ/PVctwl5e+a2YtmdnF0WinHzP4WTVs82MymR2XvmlnvMtt97bRFye/czK4xswwz22JmE8ysQ7lwJgJ7AKdU9fsQaUiUPIgk3glAV8LiXC8Spo/dpffBzLoA04DDgeuBHxDWJelF6CK/IKp6BaG7/Kg4Y9gbeA34P+Bc4CPgTTM7Op6dmNlhwAvAHOAcwgfhPyuo9xvgccKU2KdH9+8wsyur2H0nYDNwLTAM+DNwMbueZhgH/LBsshTdP5fwjb4yQ4BWhHaXdyRh3YerCMu3nx+95l+A0cBPCdNaj61i/yXOJ6yjMYKwpsLpwF1lK3hY2XYeYfEukUZB3WQiifdjIBuY5O4FZjaZsCjPzf7VfPHXAO2Bgf7VAjVvl+zAzD6P7s5394/jDcDdHy2zryTCAlHfIsz/PzWOXd1IWB/m/Cj2N80sDbizzP7bEcZn3Onuf4iKJ1tYIvhWM3vc3YsriHEuIXEq2c9Uwqmep8zsKncvIMzf/wChJ6ckWTgfSAX+UUXcA4GN7r6+gufaAGe6+9bodY8Hfgkc5+7vR2XdgTFm1tp3XWCpvELgLHcvirY7iDDG5fJy9eYAg6vYj0iDop4HkQSKPljPBl6OPvwgfAD2JXzjLXECIbmok5UNzaxn1LW/BigifMidDOwX564GAxN910VyXipX5yggHfiXmaWU/ABTCD0wPSuJ0czsajObb2Y7oxifIyxo1RtKv7W/CFxUZtOLopg2VRH3XkBlV5LMLEkcIkuBAnYdDLk0uu1exWsAvFOSOETmA13KnnqJbIxiEmkU1PMgklinAh2AN8qc+34XyCf0SEyLyvYkLBtf66KeholAW+B3hA/CXOB2wlUc8dgL2FCurPzjTtHtvEr20QtYWUH51cB9wD3Ae8AWwmmcMUDLMvXGA++aWf/o8bGE1WKr0pLwO69I+UGPBcB2d4+VKyvZT1Uq2pcBLcrsgyiW6vYl0mAoeRBJrJKxDf+q4LnzzeyaqAt/E9CtBvvPi27Lf7PtWOb+PsAAwrLDk0oKzaxVDV5vHV9POMo/Lrma4XSgotMEiyrZ9w+Bf7n7LWViPKh8JXd/38yWEMYpGJAJ/K+auDcTkriGogNf/Z5EGjwlDyIJEg3kO51wmqL8YLsBhHP3Q4G3COMbfm1mXSs5L1/ZN9+M6PZAorEL0XwG7crUKUkSSr95xYeDRgAAAjBJREFUm1kf4Gjgc+IzAzjDzG4qc+rinHJ1pgE7ge7u/p849t2Kr/cOXFBRRcJVKyXjCJ6taAxFOYuA7maW5u6V9UAkUl/C2BGRRkHJg0jinAm0Bka7+ydln4gGA95C6Jl4C3gQuBD4wMz+CKwmJATp7n4vsIrwgfwzM9sKFEYTDk0H1gAPm9lvCZcAjgK2lXm5hYQk4/6oTlvgD9F28foTYa6Kf5rZeODbhEGXpdw928xuA0ZHScr7hPFW+wFD3f3sSvY9mZBAfQIsIyQOlc3G+QxhkGYK8PRuxD2VMKjyYCChs1pWYhDhdynSKGjApEji/BhYUj5xAHD3QsIljudE34azCD0BnwIPAa8TLvdbFdXPI1wBMJAwHmBGVF5AGJAZIwwkvA74FWG8QMlr5RN6B4qiOncAd0f7iUuUsAwn9Jy8ApwF/KiCevdG8Z8KvErofbkA+KCK3d8e1bszui0gTOxUURzrCEnMVHev7DRI2fqLgS+ieOqVmQ0AOvP1gaYiDZbtOkhaRKTxMbM9CD0nV7r7+N3c5hrgEnf/dp0GV30cdwOHu7vmeZBGQ8mDiDRa0UyRBwEjCZMs9a1m3oWy26YDy4EL3P2tuouy2hhWAue5+7v1EYNITWjMg4g0ZgMJE1ytBC7c3cQBwN1zzexnhDko6ktv4HYlDtLYqOdBRERE4qIBkyIiIhIXJQ8iIiISFyUPIiIiEhclDyIiIhIXJQ8iIiISFyUPIiIiEpf/Bxv7yc3bW1HyAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(8,5))\n",
    "ax.plot(delays, mean_mse_delay, label='Mean MSE linear regression', color='blue', linewidth=5)\n",
    "ax.plot(delays, mean_mse_delay_dum, label='Mean MSE dummy mean', color='orange', linewidth=5)\n",
    "ax.legend(fontsize=15)\n",
    "ax.set_xlabel(\"Actual delay (min)\", fontsize=15)\n",
    "ax.set_ylabel(\"Mean MSE\", fontsize=15)\n",
    "ax.set_xlim((0,140));\n",
    "ax.legend(fontsize=15);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 5) Ridge Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's look at a Ridge Regression and see how it manages to reduce the influence of some features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[====================================================================================================] 100%"
     ]
    }
   ],
   "source": [
    "n_alphas = 10\n",
    "alphas = np.logspace(1, 10, n_alphas)\n",
    "ridge = linear_model.Ridge()\n",
    "\n",
    "coefs = []\n",
    "errors = []\n",
    "count=0\n",
    "for a in alphas:\n",
    "    ridge.set_params(alpha=a)\n",
    "    ridge.fit(X_train, y_train)\n",
    "    coefs.append(ridge.coef_)\n",
    "    mse = metrics.mean_squared_error(y_test, ridge.predict(X_test))\n",
    "    rmse = math.sqrt(mse)\n",
    "    mae = metrics.mean_absolute_error(y_test, ridge.predict(X_test))\n",
    "    r2 = metrics.r2_score(y_test, ridge.predict(X_test))\n",
    "    errors.append([mae, rmse, r2])\n",
    "    display_progress_bar(len(alphas),count)\n",
    "    count+=1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, (ax1, ax2) = plt.subplots(1, 2, sharey=False, figsize=(10,5))\n",
    "\n",
    "ax1.plot(alphas, coefs)\n",
    "ax1.set_xscale('log')\n",
    "ax1.set_xlabel('alpha', fontsize=15)\n",
    "ax1.set_ylabel('weights', fontsize=15)\n",
    "ax1.set_title('Ridge coefficients as a function of the regularization', fontsize=15)\n",
    "ax1.axis('tight')\n",
    "\n",
    "ax2.plot(alphas, errors)\n",
    "ax2.set_xscale('log')\n",
    "ax2.set_xlabel('alpha', fontsize=15)\n",
    "ax2.set_ylabel('error', fontsize=15)\n",
    "plt.axis('tight');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The Ridge regularisation has reduced the weight of nearly every feature except for the departure time of day in seconds (which is still reduced to 0.001). The outcome is not better than our linear regression baseline, and regularisation even tends to worsen the results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index([0, 'DEP_TIME_SECS_STD', 'ARR_TIME_SECS_STD', 'DISTANCE_STD'], dtype='object')"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.iloc[:,[0,214,215,216]].columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['DEP_TIME_SECS_STD', 'ARR_TIME_SECS_STD', 0, 'DISTANCE_STD', 2], dtype='object')"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_coefs = coefs[-1]\n",
    "final_coefs[(-final_coefs).argsort()][:5]\n",
    "X_train.iloc[:,(-final_coefs).argsort()[:5]].columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[12.046983227204498, 19.729632178618054, 0.012743733795572476],\n",
       " [12.046919170243596, 19.72938346915575, 0.012768624115400029],\n",
       " [12.047272026224814, 19.72841517513859, 0.012865525955071733],\n",
       " [12.04856829922698, 19.72552358952672, 0.01315487252790759],\n",
       " [12.053449119337737, 19.71929632463499, 0.0137778598932915],\n",
       " [12.087214383300836, 19.72366365126565, 0.013340964873478023],\n",
       " [12.185059229749118, 19.771217428830518, 0.0085775571778135],\n",
       " [12.260706664201301, 19.851363024989087, 0.0005235068126338005],\n",
       " [12.273256275654514, 19.86706805456484, -0.0010585525131689977],\n",
       " [12.274643395868017, 19.86877577981081, -0.001230657066472185]]"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "errors"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 6) Lasso Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's look at a Lasso Regression and see how it manages to set the influence of features to 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[====================================================================================================] 100%"
     ]
    }
   ],
   "source": [
    "n_alphas = 4\n",
    "alphas = np.logspace(-1, 2, n_alphas)\n",
    "lasso = linear_model.Lasso()\n",
    "\n",
    "coefs_l = []\n",
    "errors_l = []\n",
    "count=0\n",
    "for a in alphas:\n",
    "    lasso.set_params(alpha=a)\n",
    "    lasso.fit(X_train, y_train)\n",
    "    coefs_l.append(lasso.coef_)\n",
    "    mse = metrics.mean_squared_error(y_test, lasso.predict(X_test))\n",
    "    rmse = math.sqrt(mse)\n",
    "    mae = metrics.mean_absolute_error(y_test, lasso.predict(X_test))\n",
    "    r2 = metrics.r2_score(y_test, lasso.predict(X_test))\n",
    "    errors_l.append([mae, rmse, r2])\n",
    "    display_progress_bar(len(alphas),count)\n",
    "    count+=1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnIAAAFVCAYAAABrbwKGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xd4FOX2wPHvSQdCJ9RAQpceIGBBrgWR5sVeUBERRcReQO/Vn3q999qwXBsiCqioCCp2VEBEigoJvfceeu+Q5Pz+mIkuYVM27GY3yfk8zzzZmXln5mx2s3ty5p13RFUxxhhjjDFFT1iwAzDGGGOMMQVjiZwxxhhjTBFliZwxxhhjTBFliZwxxhhjTBFliZwxxhhjTBFliZwxxhhjTBFliVwAiUg1EZkmIgdF5GVxjBKRvSIyW0Q6isiKfOznJhGZWBgxhxoRaSwi89zf4X2FeNw6InJIRMIL65jucU95z+Rzm/UickmgY/M3EblVRGacwfb/FJH3/BmTu99hIvJ//t6vMcYEQkRhH1BE1gO3q+rkwj52EPQHdgHlVFVFpCPQGYhX1cNum8Z57URVPwY+9kdAIqJAQ1Vd7Y/9FYLBwFRVbR3Ig2R/X6rqRiA2kMfMwSnvmewrReR9YLOqPlHYgYUaVX32TPchIrfivO7ne+x3wJnu1xhjCotV5AIrAVjq8YWcAKz3SOJM3hKAJcEOohBlf88EjYgU+j96+RXKsRljTKFS1UKdgPXAJV6WVwS+A3YCe93H8R7rbwXWAgeBdcBN7vIGwK/AfpxKxliPbc4DUtx1KcB5ucRVGxjvHn838Ka7PAx4AtgA7AA+BMp7bHcO8BuwD1gAXOgufx84CZwADgF3AseADHf+X8CFONWVvGK4FZjh0e4sYBKwB1gBXOex7n3gLeB793c1C6jvrpsGKHDYjeF6oIr7u97n7m86EJbD7+g1YBNwAJgDdPRY1x5IdddtB17JYR+5vs7Z2k5xf1/H3HgbAVNxKijk8LtRYACwyt3/W4B4rL8DWOb+bpYCbYDRQCZw1D3OYCDR3VeEu11N4Bv3d7QauMNjn08D49z3xkGcxDM5l/ea1/ell/fMJdm2659t/bcef1OPAAvdfY4FYjy2uwyY777GvwEtc4lNgbvd39+6fLzfKgPfuq97CvCfrNcj++/QXfbn6+fltcvt/fU08Dnwkbv+dnfZR+76N93fSdaUDjztrnsMWOPxml/pLm/CqX+T+zxeh/9ke8+sdp//N0DN/L7fbLLJJpsCPRX+AXNO5CoDVwOlgbLAZ8BX7roy7od3Y3e+BtDMfTwGeBwn4YoBzneXV3I/WHvjnELu5c5X9nLscJwk7FX3WJ77uc39EK+Hc6ptPDDaXVcLJ+Hq7h6/szsf567P/oWQ/YvrQtxELo8Y/tzOXbcJ6Os+rzY4CWwzj2PuwUmsInBOyX7qcUwFGnjMPwcMAyLdqWNOX0TAze7rFAE8DGzDTRiA34He7uNY4Jwc9pHj65xD+6mcmrhln8/+O1Wc5LACUAcnYezqrrsW2AK0AwTnn4AEb+9LTk/kfgWGuq9LkrvfTu66p3ESgu7u6/gc8EcOzyfX92X294yX7U9b78Y+GyfZrISTqA5w17XB+QfkbDe2Pm776Bz2rzhJWyWgFHm/3z51p9JAU7dtQRO53N5fT+MksVfg/K2VwiORy/Ycsl6f1h6ve013u+tx/pGp4S2G7L9j4GL3+bYBooE3gGn5eb/ZZJNNNhXGFDKnVlV1t6p+oapHVPUg8F/gAo8mmUBzESmlqltVNet020mc01E1VfWYqmZ1nu4BrFLV0aqarqpjgOXA370cvj3OB/0gVT2cbT834VSX1qrqIeAfwA3uqZ2bgQmqOkFVM1V1Ek5VqnsBfgW5xeDpMpzTs6Pc5zUX+AK4xqPNeFWdrarpOIlcUi7HPYmTGCeo6klVna6qXk/rqepH7uuUrqov43yxNfbYTwMRqaKqh1T1jxz2kdfr7A/Pq+o+dfq5/cJfz/924EVVTVHHalXdkNfORKQ2cD7wqPu6zAfew0nGssxw3wcZOBW+Vjnszpf3pS9eV9U0Vd2DUyHLes53AO+o6ixVzVDVD4DjOJXknDynqntU9Si5vN/cC0GuBp5yX8+lwAcFfQJ5vL8AflfVr9y/taPe9iEiccBXwL2qOs/d72fu7yZTVcfiVM/a5zOsm4CRqjpXVY/j/P2fKyKJHm1yer8ZY0zAhUwiJyKlReQdEdkgIgdwTgNWEJFwdfqUXY9zCmOriHwvIme5mw7Gqa7MFpElInKbu7wmzulQTxtwqmjZ1QY2uIlPdtn3swGnYlANJ4G8VkT2ZU04X/g1fHz6ecXgKQE4O9sxbwKqe7TZ5vH4CLl32h+CU3GcKCJrReSxnBqKyMMiskxE9rvHLY9zahagH86pz+UikiIil+Wwjxxf59yftk9yev61cU6x+aomsMdNPLNkfy9lP2ZMDv24fHlf+iKn55wAPJzt/VLbjSMnmzwe5/Z+i8P5W9iUw7Y+yeP9lee+RSQS5/TrJ6r6qcfyW0Rkvkf8zbPtNzenvF7uP3O7yf21D8ZFMsaYEiqUOgw/jPPf99mquk1EkoB5OEkaqvoT8JOIlMLph/MuTh+abThVB0TkfGCyiEwD0nC+hDzVAX70cuxNQB0RifCSSGXfTx2c/jfb3e1Gq+odBXzO+Y0he7tfVbWzH46Jm5w8jPNl3wz4RURSVPVnz3buFbePAp2AJaqaKSJ7+ev1WQX0EpEw4CrgcxGprKdf2JHr65wPh3FO42WpnlNDLzYB9XNYl9vFBWlAJREp65HM1cE5TesrX96X3vh6EcQm4L+q+l8ftvE8Ro7vNzf5TgfigZXu4toeTbJe+9I4XSMgh9crr/eXl7i8eQOnH9yfV/SKSALOZ0UnnIpehojM99hvXvs85fUSkTI4p38L8tobY4zfBasiFykiMR5TBE5/qaPAPhGpBDyV1dgdW6un+yF6HKdjcoa77loRiXeb7sX5YM4AJgCNRORGEYkQketx+vB85yWe2cBW4HkRKePG1MFdNwZ4UETqikgs8CzOBRXpOB2v/y4iXUQk3N3uQo94fJFbDJ6+c59XbxGJdKd2ItIkn8fZjtPfDwARuUxEGoiI4HzZZrhTdmVxvrR3AhEi8iRQzmM/N4tInKpm4nSqJ5f9eH2d82k+cJVb2WuAUwnMr/eAR0SkrTgauF/0kO334klVN+FcJPCc+7q0dI9bkCFhfHlfepNjnDl4FxggIme7z7mMiPQQkbL53D7H95t7Gnk88LT7epwF3JK1oaruxEl4bnb/Pm4j50Q61/dXXkTkTpxT9De678EsZXA+E3a67friVOSybAfiRSQqh11/AvQVkSQRicb5+5+lquvzG5sxxgRSsBK5CThf5lnT08D/cDow7wL+4NQKRRhOJScNpyP/BcBAd107YJaIHMK5oux+VV2nqrtx+vc8jHMqZDBwmaruyh6M+4X0d5zO7xuBzTincgFG4vR5moZztewx4F53u03A5cA/cb4oNgGDKMDvNY8YPNsdBC4FbnB/H9uAF3D6E+XH08AH7mmm64CGwGSc5Ph3YKiqTvWy3U/ADziVlw04vwfPU11dgSXu6/AacIOqHvOyn9xe5/x4Feeqze04/bHynUyp6mc4ffI+wancfIXTqR+cCxSecH8vj3jZvBdO5/004EucfmGTfIwdX96XORgBNHXj/Cofx0vFqVi/ifOPzmqcDv75jTev99s9OKdAt+H8nYzB+Wcryx04fxO7gWY4CbE3eb2/8tILJ8FNE2cg50Mi8k+3397LOO/t7UALYKbHdlNwrjLeJiLePht+Bv4Pp1/gVpxE9AYf4jLGmICSHPq1G2OMz0TkBaC6qvYJdizGGFMShMzFDsaYokdEzhKRlu5p2/Y4p5y/DHZcxhhTUoTSxQ7GmKKnLM7p1Jo449W9DHwd1IiMMaYECalTq+KM1/UhzpVtmcBwVX0tW5sLcb4o1rmLxqvqM4UZpzHGGGNMKAi1ilw68LCqznWvqpsjIpPcDsuepquq13HKjDHGGGNKipBK5FR1K86VYajqQRFZhjPwZvZEzidVqlTRxMTEMw/QGFNkzJkzZ5eqxgU7Dn+wzzBjShZfPr9CKpHzJM4tcFrj3PQ9u3NFZAHOcAiP6F+36/Lcvj/OTcapU6cOqampgQvWGBNyRCTP268VFYmJifYZZkwJ4svnV0heteoOvPsF8ICqHsi2ei7OfUFb4Yzk7nUsLVUdrqrJqpocF1cs/ik3xhhjjDlFyCVy4twv8QvgY1Udn329qh5w73eIqk7AuUtEfu+baIwxxhhTbIRUIufeJmoEsExVX8mhTXW3He64VWE4o8YbY4wxxpQoodZHrgPQG1jk3tganNtf1QFQ1WHANcBdIpKOc3uvGzSUxlAxxhhjjCkkIZXIqeoMQPJo8ybOfSONMcYYY0q0kDq1aowxxhhj8s8SOWOM8YGI1BaRX0RkmYgsEZH73eWVRGSSiKxyf1bMYfs+bptVItKncKM3xhQ3lsgZY4xvsu5A0wQ4B7hbRJoCjwE/q2pD4Gd3/hQiUgl4CjgbaA88lVPCZ4wx+WGJnDHG+EBVt6rqXPfxQSDrDjSXAx+4zT4ArvCyeRdgkqruUdW9wCSga+CjNsYUVyF1sUMo2LX5EGXKR1GqbFSwQzHGhLhsd6Cp5t5mEFXdKiJVvWxSC9jkMb/ZXeZt36fcncaYQqMKGSch44QzZaa7j0/+tVwzgh2ld6E8iEVMeahU1++7tUTOg6oyedQSDu09TrvL6tL8glqEh1vR0hhzuux3oHGHt8xzMy/LvH7zqOpwYDhAcnJyCH87mRypnp4EZZ48PSn6s80JyHAfZ3qs96ldQY7h2e6k09b4X5OecP1ov+/WEjkPIsKl/ZozfdxKZoxbxdIZaZx/XUNqn1Up2KEZY0JIDneg2S4iNdxqXA1gh5dNNwMXeszHA1P9FtiG32HJaTfE4bT80WvS6a823jbL3s7LdgFr46Vd9jaa6SUJ8kh8vFWkMrMlQN62zTzpJRY/CouA8CgIi4TwSOdxuLdl7uPI0qfOh0f9tY+8tg3LNi/h+X/9C12IxlW2WkB2a4lcNpVqlqHn/UmsW7CLmZ+v4pv/zade6zg6XN2AclVKBTs8Y0yQ5XIHmm+APsDz7s+vvWz+E/CsxwUOlwL/8Ftwe9bAos9OXXbaqSYvxb3TFnlrk5/9eCscavDaeG3nbTvJIWmJ+Gu5ZyITWS5bEuSZGEWeOu8tCcqznUdylf0YnvsK2UTKFCZL5LwQEeolxVGnWSXmT9rEnB/Xs2Hxblp3rkObrglERoUHO0RjTPDkdAea54FxItIP2AhcCyAiycAAVb1dVfeIyL+BFHe7Z1R1j98ia32zMxljSgwpCXe3Sk5O1tTU1AJvf2jvMX4bv4ZVKduJrRjNeVc3oEHbquSzT4wxJghEZI6qJgc7Dn84088wY0zR4svnl/Xkz4fYijFc2q8ZVz7chpjYSCa+t4SvXpnHrs0Hgx2aMcYYY0owS+R8ULNhBa79RzsuuLExe9IOM+6/Kfz6yQqOHQpwh1ZjjDHGGC+sj5yPwsKE5n+rRYO2VZn93ToW/7qFVanbObtnPZp1rEmYDVdijDHGmEJiWUcBxZSJ5G/XN+L6x9tRpXZZpn26knHPprB5xd5gh2aMMcaYEsISuTNUuVYslz+QRNf+zTlxNIOvX53Hj8MXcWD30WCHZowxxphizk6t+oGIUL9NVRKaV2bepI3M/XED6xftps2ldWjdxYYrMcYYY0xgWCLnRxFR4bTrUZezzq3Bb1+sJuX79Sz7fSsdrm5I/TZxNlyJMcYYY/zKTq0GQNlKMXS5ozlXPNSa6FKR/PTuYr5+dR67txwKdmjGGGOMKUYskQugWo0qct0/k7mgVyN2bTnE2P/MZtqYFRw7bMOVGGOMMebM2anVAAsLD6P5BfE0aFuNWd+uZfG0LaxM3c45PevRtGMtwsLsdKsxxhhjCsYqcoUkJjaSC3o15rrH21OlViy/jnGGK0lbZcOVGGOMMaZgLJErZFXiY7n8wdZ0uaM5xw+f5MuX5/HTe4s5uOdYsEMzxhhjTBFjp1aDQERo0LYqCS0qM++nDcyduJH1C3bRpmsCrTvXIcKGKzHGGGNMPlgiF0SRUeG0/3s9Z7iS8auZ/e06lv22lQ7XNKBekg1XYowxxpjc2anVEFCuSim69m/B5Q+2JjI6nB/fWcw3r81nd5oNV2KMMcaYnFkiF0LiG1fk+sfb0fH6RuzceJCx/0lh+tiVNlyJMcYYY7yyU6shJiw8jJYXxdOwXVVmfbOOhVM3szJlO+dcXo8mHWracCXGGGOM+ZNV5EJUqdgoLryxMdf9sx0Vq5dm6scr+Oy5FLau3hfs0IwxxhgTIiyRC3Fxtcty5cNtuLRfM44dOsn4l+YyccQSDu214UqMMcaYks5OrRYBIkLDdtVIbFmFuT9tYN7EjaxbsJO23RJJuqQ2EZE2XIkxxhhTElkiV4RERodzds96NDmvBjM/X82sr9eybGYaHa5pSN1WVWy4EmOMMaaEsVOrRVC5KqXoNqAFPe9PIjwynB+GLeLbNxawZ+vhYIdmjDHGmEJkiVwRVrtJJa5/oh3nX9uQ7esOMPbfs5kxbhXHj9hwJcYEkoiMFJEdIrLYY9lYEZnvTutFZH4O264XkUVuu9TCi9oYUxyF1KlVEakNfAhUBzKB4ar6WrY2ArwGdAeOALeq6tzCjjVUhIeH0apTbRq1r8Yf36xlwS+bWJmyjXOuqE+Tc2sgNlyJMYHwPvAmzucVAKp6fdZjEXkZ2J/L9hep6q6ARWeMKTFCrSKXDjysqk2Ac4C7RaRptjbdgIbu1B94u3BDDE2lykZx0U1ncd0/2lE+rjS/jF7OZ8+nsnVNbt8lxpiCUNVpwB5v69x/Nq8DxhRqUMaYEimkEjlV3ZpVXVPVg8AyoFa2ZpcDH6rjD6CCiNQo5FBDVlydslw1qA2db2vKkf3HGT9kDpNGLeHwvuPBDs2YkqIjsF1VV+WwXoGJIjJHRPrntBMR6S8iqSKSunPnzoAEaowp+kLq1KonEUkEWgOzsq2qBWzymN/sLtuabfv+OBU76tSpE6gwQ5KI0Kh9dRJbVmHOjxuYP3kja+fvol33RFpdXJvwyJDK340pbnqRezWug6qmiUhVYJKILHcrfKdQ1eHAcIDk5GQNTKjGmKIuJL/RRSQW+AJ4QFUPZF/tZZPTPuRUdbiqJqtqclxcXCDCDHlRMRGce0V9bnzqbOIbV+T3L9cw5plZrFu4C1X7XjDG30QkArgKGJtTG1VNc3/uAL4E2hdOdMaY4ijkEjkRicRJ4j5W1fFemmwGanvMxwNphRFbUVU+rjQ9Brbk7/e1IixcmDB0Id+9uYC922y4EmP87BJguapu9rZSRMqISNmsx8ClwGJvbY0xJj9CKpFzOwmPAJap6is5NPsGuEUc5wD7VXVrDm2NhzpNK3P9/7WnwzUN2LZmP58+M5uZn6/ixNH0YIdmTJEiImOA34HGIrJZRPq5q24g22lVEakpIhPc2WrADBFZAMwGvlfVHwsrbmNM8RNqfeQ6AL2BRR5jMP0TqAOgqsOACThDj6zGGX6kbxDiLLLCw8NIuqQOjdpX54+v1zD/502smL2dc6+ox1nn2HAlxuSHqvbKYfmtXpal4XxmoaprgVYBDc4YU6KEVCKnqjPw3gfOs40CdxdORMVX6XJRXNy7Cc3/VovpY1cy5cPlLJ6WRsfrG1K9bvlgh2eMMcaYfAipU6um8FVNKMdVj7TlklubcGjvMb54YQ4/v7+Uw/ttuBJjjDEm1IVURc4Eh4QJjc+pQd2kOOb8sJ75kzexZt5Oknu4w5VEWL5vjDHGhCJL5MyfomIiOPfKBjQ5ryYzP1/F7+PXsGzmVjpc04DEFlWCHZ4xxhhjsrFSizlNhWql6XF3Ky67x+mT/f1bC/nurQXs234kyJEZY4wxxpNV5EyOEppXJv6siiz8ZTMp369jzDOzaNWpNsndE4mKsbeOMcYYE2z2bWxyFR4RRuvOdWjUvhp/fLWGeRM3suKPbZx7VX0at69uw5UYY4wxQWSnVk2+lCkfTac+Tbn60bbEVorh5/eX8cWQOWxfn/0OasYYY4wpLJbIGZ9Ur1ueawa35eJbmnBg9zE+fz6Vnz9cxrFDJ4MdmjHGGFPi2KlV4zMJE5qcV4P6reNImbCehVM2cezQSXoMbBns0IwxxpgSxRI5U2BRpSLocHUDoktFMOubtezYcICqCeWCHZYxxhhTYtipVXPGWl4UT3TpCFK+Xx/sUIwxxpgSxRI5c8aiSkWQdElt1i/cxc6NB4MdjjHGGFNiWCJn/KLFRbXdqty6YIdijDHGlBiWyBm/iC4VQatOtVm3wKpyxhhjTGGxRM74TcuLrSpnjDHGFCZL5IzfnFKV22RVOWOMMSbQLJEzftXyoniiSkWQalewGmOMMQFniZzxq+jSkbTqVJu183eya7NV5YwxxphAskTO+F2ri52qnI0rZ4wxxgSWJXLG76JLR9Lq4njWztvJrs2Hgh2OMcYYU2xZImcCouXFtYmKCSfVrmA1xhhjAsYSORMQMWUiadmpNmusKmeKIREZKSI7RGSxx7KnRWSLiMx3p+45bNtVRFaIyGoReazwojbGFEeWyJmAaZVVlZtgVTlT7LwPdPWy/FVVTXKnCdlXikg48BbQDWgK9BKRpgGN1BhTrFkiZwImpkwkLS+uzZq5O9m9xapypvhQ1WnAngJs2h5YraprVfUE8ClwuV+DM8aUKJbImYBq1ak2kTHhdrcHU1LcIyIL3VOvFb2srwVs8pjf7C47jYj0F5FUEUnduXNnIGI1xhQDlsiZgIopE0krq8qZkuFtoD6QBGwFXvbSRrwsU287U9XhqpqsqslxcXH+i9IYU6xYIudFemZ6sEMoVv6qyq0PdijGBIyqblfVDFXNBN7FOY2a3Wagtsd8PJBWGPEZY4onS+Q8qCqDfh3Ef/74T7BDKVZiykTS8qJ41szbYVU5U2yJSA2P2SuBxV6apQANRaSuiEQBNwDfFEZ8xpjiyRI5DyJC1dJVGb9qPMv3LA92OMVKUqc6REaFkzphfbBDMeaMicgY4HegsYhsFpF+wIsiskhEFgIXAQ+6bWuKyAQAVU0H7gF+ApYB41R1SVCehDGmWLBELps7W91J+ejyvJjyIqpeu66YAoiJdapyq+fuYHeaVeVM0aaqvVS1hqpGqmq8qo5Q1d6q2kJVW6pqT1Xd6rZNU9XuHttOUNVGqlpfVf8bvGdhjCkOLJHLplxUOe5JuoeUbSlM2TQl2OEUK0mXWFXOGGOM8SdL5Ly4utHVNKjQgJdTX+ZExolgh1NsxMRG0uKieFbP2cGetMPBDscYY4wp8iyR8yIiLIJB7Qax6eAmPln2SbDDKVaSLqntVuVsXDljjDHmTIVcIuftHobZ1l8oIvs97mf4ZCDiOK/meVwQfwHvLHyH3Ud3B+IQJVKp2ChaXBjPqjk72LPVqnLGGGPMmQi5RI6c72HoabrH/QyfCVQgDyc/zLH0Y7w1/61AHaJESupcmwjrK2eMMcacsZBL5M7gHoZ+V7d8XW446wa+WPUFK/asCHY4xUap2ChaXliLVanb2bvNqnLGGGNMQYVcIpdP54rIAhH5QUSaeWvgr/sUDmg1gLJRZRmSOsSGI/GjpEvqEBFld3swxhhjzkRRTOTmAgmq2gp4A/jKWyN/3aewfHR5BrYayKyts5i6aWqB92NOVapsFC0uqMVqq8oZY4wxBVbkEjlVPaCqh9zHE4BIEakSyGNe2/ha6pWvx0upL3Ey42QgD1WitO5ch/DIMOsrZ4wxxhRQkUvkRKS6iIj7uD3OcwjoZaWRYZEMajeIjQc38slyG47EX5yqXDyrUqwqZ4wxxhREyCVy3u5hKCIDRGSA2+QaYLGILABeB27QQui8dn6t8zm/1vm8s+Ad9hwLiWsxioWkrKrcD+uDHYoxxhhT5IRcIpfDPQyHqeowd/2bqtpMVVup6jmq+lthxTYoeRBH0o8wdP7QwjpksVe6XBTNL4hn1ezt7Nt+JNjhGGOMMUVKyCVyoaxehXpc3/h6Plv5Gav2rgp2OMVG6851CI+wvnLGGGOMryyR89Fdre4iNjKWISk2HIm/OFW5Wqycvc2qcsYYY4wPLJHzUYWYCgxMGsjvW39n+pbpwQ6n2Gh9aYJTlbO+csYYY0y+WSJXANc1vo7EcokMSRnCyUwbjsQfSpeLotkFtVg5y6pyxhhjTH5ZIlcAWcORrD+wnrHLxwY7nGIjq6/cHKvKGWOMMfliiVwBdazVkfNqnsfQBUPZd2xfsMMpFsqUj6bZ32qxYvZ29u2wqpwxxhiTF0vkCkhEGJQ8iMMnDzN0gQ1H4i+tL61DWLhYVc4YY4zJB0vkzkCDig24ttG1jFsxjjX71gQ7nGKhTPlomnesxYpZ29m/06pyxhhjTG4skTtDdyfdTemI0gxJHRLsUIqN1l2cqlzqDxuCHYoxxhgT0iyRO0MVYyoyoNUAZm6ZyfTNNhyJP5QpH02zjjVZ8cc2q8oZY4wxubBEzg96ndWLhHIJDEm14Uj8pU2XBLevnFXlTOgRkZEiskNEFnssGyIiy0VkoYh8KSIVcth2vYgsEpH5IpJaeFEbY4ojS+T8IDI8kkeSH2Hd/nWMWzEu2OEUC1lVueV/bGP/zqPBDseY7N4HumZbNglorqotgZXAP3LZ/iJVTVLV5ADFZ4wpISyR85ML4i/gnBrn8PaCt9l/fH+wwykW2lyaQFiYMOfH9cEOxZhTqOo0YE+2ZRNVNd2d/QOIL/TAjDEljiVyfiIiDGo3iIMnDjJswbBgh1MslKng9pX7fRsHdllVzhQptwE/5LBOgYkiMkdE+ue0AxHpLyKpIpK6c+fOgARpjCn6LJHzo0YVG3FNw2v4dPmnrN2/NtjhFAttuiQgYTaunCk6RORxIB34OIcmHVS1DdANuFtE/uatkaoOV9VkVU2Oi4sLULTGmKIuItgBFDd3t76bCesm8HLqy7zV6a1gh1PklakQTdOONVnczKkgAAAgAElEQVTy6xbadkukXJVSwQ7JmByJSB/gMqCTqqq3Nqqa5v7cISJfAu2BaYUXpSlOVJX0TOVEeiYn0jM5mZHJcffniYzsy05tdyL9rzYnMjI5mTWftcyj3ckM5bhHu4xMr2/vgj0H/Lgv/+3K2Z8f93VOvUoM6nKWH/fosETOzyrFVGJAqwG8lPoSM7fMpEOtDsEOqchrc2kCS6ZvYc6PG7joZv//ERjjDyLSFXgUuEBVvY6bIyJlgDBVPeg+vhR4phDDNAWgqmRkqpvEKMczMv5MbjwTpVMTHyfpOZ59WXpW4qSnJ1Re2nlPzPSUJMyfRCAqPMyZIsKIdH+e8jhcCA8TBPHXURF/7Qr8ui/Ab88zKjzcL/vJzhK5ALjxrBsZt2IcQ1KGcHaNs4kIs1/zmYitGE2z82uxZNoW2nZNsKqcCToRGQNcCFQRkc3AUzhXqUYDk8T5JvlDVQeISE3gPVXtDlQDvnTXRwCfqOqP/opr+4FjrN15GFUlU51KR6ZCpiqqiiqnzGcq7jJ1l7nbZHLq/J9tAD11/q/9/nVMVcjMPDUG9TiGt/k/YyTnfWc6AZ0y74TkEXPWvOKx/1x+AumZ+lc1Kodqlb8rPVGnJEjyV6IUHka0R9JUJjriz3ZZCVZkhBAVHv5nUuUt4Yr22N+pSdipxzx1WRgRYYL4OxMyAWUZRgBEhkfyUPJDPPDLA3y+8nNuOOuGYIdU5LXpUoclM6wqZ0KDqvbysnhEDm3TgO7u47VAq0DFNXHpdv7vq8V5NwyCMHEuCjvlJ87PMHEqMlnLnXlnWdZ8mJtchIX9NS/gtpEc9pFtnr+Wh4cJEe52YWHiJjPOT8+kyFsiFBkRRnS2hOrPxCh7Beu0BMppa8mS8Zd8J3IicjVQQVVHuPN1cTrzNgV+Bvqp6r6ARFkEXVz7YtpXb89b89+iW91ulI8uH+yQirTYijE061CTJdPTaNstgXKVrSpnTHadm1SjQVys96Qpe4KDeCRFgEdS9VcS9Nd8mADZ5gVBPPYhiMc2px7bGBMYvlTkngA+9Jh/A6gCPA/cCfwXuNt/oRVtIsLgdoO59ttreWfhOwxuNzjYIRV5bbomsGRmmlOVu8mqcsZkV718DNXLxwQ7DGNMIfJl+JF6wCIAESmP00n3QVV9Hngc+Lv/wyvaGldqzFUNr2LMsjGs378+2OEUebEVY2jaoSbLf9vKgd02rpwxxhjj6zhyWd09LwAygMnu/GbABjry4p7W9xAdEc3LqS8HO5RioU2XBADm/mj3YDXGGGN8SeQWADe5l8zfDvyiqsfddXWAHf4OrjioUqoK/Vv2Z+rmqfyW9luwwynyylZyqnLLftvKwT3Hgh2OMcYYE1S+JHL/BK4EDuBU5P7lse4KYJYf4ypWbm5yM/Gx8QxJGUJ6ZnreG5hctenqVOXmWFXOGGNMCZfvRE5VZ+BU3toDCarqmbiNxOknZ7yICo/i4eSHWb1vNeNXjQ92OEVe2UoxNOlQk2Uz06wqZ4wxpkTLdyInIk8CZVV1jpdhRuYB3sZVMq5OdTqRXC2ZN+e9ycETB4MdTpHXtqv1lTPGGGN8ObX6FBCfw7qa7nqTg6zhSPYd38fwhcODHU6RV7ZSDE3Oq8HS36wqZ4wxpuTyJZETcr5/bDyw98zDKd6aVG7ClQ2v5KNlH7HxwMZgh1PktemaAApzf7KqnDHGmJIp10RORPqIyBQRmYKTxL2dNe8x/QZ8BPxaGAEXdfe2vpeosCgbjsQPylUuxVnn1WDpzDQO7bWqnMmdiESKSAf33qfGGFMs5FWROwLsdicB9nvMZ03rgBeB/oELs/ioUqoKd7S8gymbpjBrq13oe6badkmATOsrZ/IlA5gCNAl2IMYY4y+53qJLVT8DPgMQkVHAv92bPpsz0Ltpbz5b8RkvprzIuMvGER4WHuyQiqxyVZyq3JKZabTpmkhsxehgh2RClKpmisgqoFqwYzHGGH/xZfiRvpbE+Ud0eDQPJT/Eyr0r+XL1l8EOp8hr29WtyllfOZO3x4EnRaRFsAMxxhh/yLUil52IJANX4VzccNqdmVX1ujMNSERGApcBO1S1uZf1ArwGdMc59Xurqs490+MWtksTLqVN1Ta8Me8NuiR2oWxU2WCHVGSVq1KKs86tztIZabTpkmBVOZObJ4DKwHwR2QJsJ9tFXKraPhiBGWNMQfgyjtxdwGyc23PVx7m3avbJH94HuuayvhvQ0J36A2/76biFSkQY3H4we4/t5d1F7wY7nCKvbbdENFOZO9GqciZXi4HvgA+Bn935JdkmY4wpMnypyD2CcweHAaoasPtMqeo0EUnMpcnlwIeqqsAfIlJBRGqo6tZAxRQozSo3o2f9nny09COubXgttcvVDnZIRVa5KqVofG51lk5Po22XBMpUsKqcOZ2q9g12DMYY40++jCNXFRgTyCQun2oBmzzmN7vLTiEi/UUkVURSd+7cme+df7cwjR8XbzvzKPPpvjb3EREWwStzXim0YxZXbbu6VTnrK2fyQUSqiEhDEakc7FiMMaagfEnkfgDODlQgPhAvy04bqFhVh6tqsqomx8Xl76xvZqYyauZ67vt0Hr+t2XWmceZL1dJVub3F7UzeOJmUbSmFcsziqnxcKRqfU50l09M4vO94sMMxIUpErheRZTj945YDO0RkmYhcG+TQjDHGZ3kNCNw0awLeAvqIyFMicp7nOo82hWEz4HkOMh5I88eOw8KEEX2SSahUmv4fzmHR5v3+2G2ebml6CzXK1ODFlBfJyMwolGMWV227JZJpfeVMDkSkFzAGWAv0xbloqq87/6mI3BDE8Iwxxmd5VeQWA4vcaQrOBQZPAdM9li/yaFcYvgFuEcc5wH5/9o+rUDqK0f3OpnypSG4dNZu1Ow/5a9c5iomI4aG2D7F8z3K+WfNNwI9XnJ1SldtvVTlzmseB4araQ1U/VNWf3J89gHdxrmo1xpgiI69E7iLgYo/pIi/LLvZYdsZEZAzwO9BYRDaLSD8RGSAiA9wmE3D+e16N88E70B/H9VS9fAyj+zkjEPQeMZut+4/6+xCn6ZLYhaS4JF6b+xqHTx4O+PGKs+RuCWRmKPN+svvZmtM0AL7IYd0X7npjjCkyck3kVPVXXyZ/BKSqvVS1hqpGqmq8qo5Q1WGqOsxdr6p6t6rWV9UWqprqj+NmObxvLxnp6dSLi+X9vu3Zf/Qkt4yYzd7DJ/x5mNOICI+2f5Tdx3bz3qL3Anqs4q58XGkan12NxdO3WFXOZLcdSM5hXbK7Pk8iMlJEdojIYo9llURkkoiscn9WzGHbPm6bVSLSx+dnYIwxHny52KHYU1W+eeU5Rj96H5uWLqJFfHnevSWZDXuO0Pf9FA4fD+wFu82rNKdn/Z58uORDNh/cHNBjFXdtuyU6VbmJVpUzpxgFPC0iT4jIWSJSUUQai8gTON1GRuZzP+9z+niXjwE/q2pDnDHqHsu+kYhUco9zNtAeeCqnhM8YY/LDlwGBM0UkI4cpXUT2iMgvInJlIAMOtHY9r+bk8WOM+9c/mPDGS7SoJLzRqzULN+9jwEdzOJGeGdDj39f6PsLDwnl1zqsBPU5xV6FqaRq3r8biaVaVM6d4BngJJ8laAuwClrrzL7nr86Sq04A92RZfDnzgPv4AuMLLpl2ASaq6R1X3ApPIfQB0Y4zJlS8VuYeALcAy4EVgEDAE5/L9NJzbZmUAn4vIzX6Os1CICA2Sz+bWl4dyztU3sPKPGYx68E6qrJ/Fc1c0Y/qqXTw0bj4ZmaeNduI31cpUo2/zvkzcMJE52+cE7DglQdvublVuklXljENVM1X1cZwr3y8Eerk/a6vqE+5A4wVVLevCK/dnVS9t8jUOJhR8LExjTMniSyJXE5ipqs1V9R+q+oqqPqaqzYDfgIqqegnwMTA4EMEWlsjoGDpcdzN9Xh5KzcZNmfrhuxz7/CUGt4rku4VbeeqbxZzZ533ubm12K9VKV+OF2S+QqYGtABZnFaqWplH7aiz5dQtHDgS2j6MJfSISIyITReRCVd2rqtNVdZz7c29hheFlmdcPk4KMhWmMKXl8SeT64vQv8WYU0Nt9PBZnmJIir2L1mlz12NP0fORxjh85zNGvXmegzGL8jOW8OnlVwI5bKqIUD7Z9kGV7ltlwJGcouVsiGemZzLNx5Uo8VT0GtAPCA3SI7SJSA8D9ucNLm4CNg2mMKZl8SeQigLNyWNfEY18ngGNnElQoEREatjuXvq+8zdlXXk/UxoX03fopU8d/wcjpqwN23O51u9MyriWvzX2NIyePBOw4xV2FaqVp1L46i60qZxzf4L3vmr/2nXUVah/gay9tfgIudS+yqAhc6i4zxpgC8SWR+xR4TkQecu9PWMH9+QjwLPCJ264NTr+5YiUyOobzb+jNLUPeIrFJUy7YM5OVw//NmG/9MurKaUSEwe0Gs+voLhuO5Awld3erctZXzjhJ01Ui8rmI3CYiPUSku+eUn514G+8SeB7oLCKrgM7uPCKSLCLvAajqHuDfQIo7PeMuM8aYApH89vUSkSicixvuAKI9Vh3HGZh3kKqeEJELgUP+Ht/tTCQnJ2tqqv/CUVWW/jaDb94ZStTxg1RK6sD1A++idPkKfjtGlsemP8ak9ZP49spvqRlb0+/7LykmjVrC2rk76f3f8yhdLirY4ZhCICJzVDU527K8Op2qqgbq1GuB+fszzBgT2rx9fuUk3xU5VT2hqvfj9Om4GLjR/Rmvqver6gm33dRQSuICQURo1qEjd74xnA3x57Bz/u8Mv68/8376jkw/3yv1gTYPECZhNhzJGcrqKzffqnIlXd08pnrBC80YY3zn84DA7vhHv6rqWPdniT0tUKl8WZ781yCmNe/DpvDKTBk5jI//8RBpK5f57RjVy1Snb/O+/Lj+R+btmOe3/ZY0FauXoWG7aiz6dbP1lSuhRCQauBmooKobcpqCHacxxvgi10TO7TNSzuNxrlPhhBxaqsRGM+yebsyofxUz63Tn4L69jPm/Qfw07DWOHNjvl2Pc2uxWqpauasORnKHk7olknLSqXEmlqseBxwH/94Ewxpggyasi9x1/Xan6HfCt+9Pb9G2AYgx58RVLM/r2c1gT24DP695E0y6Xs3TaFEY9cCfzJ04449OtpSNL80CbB1iyewnfr/3eT1GXPBWrl6FBslOVO3rQqnIl1CygbbCDMMYYf8krkasLzPd4XA/rW+JVo2plGdW3HVsPK68fPIsrn3mVuMR6/DxiKJ88/jBbV684o/33qNeDFlVa8L85/7PhSM5Aux6JpJ+0K1hLsMHAXSJyj4jUE5EyIlLacwp2gMYY44tcEzm3z8gJj8e5ToUTcuhqU6ciw3q3ZfWOgwyevJ2/P/YMPe4bxKG9e/jkiUeYOPyNAp9uDZMwBrcbzI6jOxi1JKdxmU1eKlYvQ8Pkaiz6dYtV5UqmWUB94HVgFXAAOJhtMsaYIiPCl8ZuZ+HbgGSc0cnvVtVVInI9sFBV/dfLv4i6oFEcL1+XxP2fzuPeMfN4++aO1GvTjt8+H8PcCV+zatZvdOzVhxYXX4qE+XatSVLVJLoldmPU4lFc1eAqasTWCNCzKN6SuyeyKnU78ydv5NwrGwQ7HFO4biOHW2IZY0xRlO9MQkQaASuB54BEoBNQ1l3dEfiHv4Mrqnq2qskzlzdn8rIdPPrFQiKiS3Fh737c8sLrVKmTwKR33+STJx5m2+qVPu/7gbYPAPC/uf/zd9glRqUaTlVu4dQtHD1kVbmSRFXfV9UPcAbjzcS5Yf1P7rKZwPhgxmeMMb7ypST0OrARJ4nrwqk3f/4VON9/YRV9vc9J4MFLGjF+7hb+O2EZqkqVOolc9+RzdL/3EQ7u3sXHTzzMpHff5OjBA/neb83YmvRp1ocJ6yYwf8f8vDcwXiV3TyT9RAbzJ20KdiimELl94sYBi4D3cO6ykDXS9rPAk8GKzRhjCsKXRK4j8Jyq7uP0UxPbATvPl819nRpw63mJjJixjqFT1wDOYMJNzr+Qvq++Q9vuPVk0ZSIjHxzAoikT0cz8DS3Sr3k/4krFMSRliA1HUkCVapShYduqLJq62apyJcurwHnAJThnFDz/IZ0AdAtGUMYYU1C+JHLHgFI5rKsF7DvzcIoXEeHJy5pyeVJNhvy0gk9m/XWlZHTp0lx4yx30fv41KteKZ+I7rzPm/waxfe3qPPdbOrI097e5n4W7FjJh3YRAPoViLbl7XU6eyGD+ZKvKlSBXAY+q6i9A9nGBNgAJhR+SMcYUnC+J3CTgnyJS3mOZuhdA3Ivz36zJJixMeOnaVlzYOI4nvlrEhEVbT1kfl1CX659+gW53P8T+ndv56J8PMnnE2xw7dCjX/f69/t9pWrkpr8551YYjKaBKNcvQoG1VFv2ymWOHTgY7HFM4SgG7c1hXltOTO2OMCWm+JHKDgDhgNTAa5/Tqkzh9TWrijJhuvIgMD+Ptm9rSuk5FHvh0PjNW7TplvYjQ9G8X0/fVYbTuehkLJ/3AyAf6s/iXSTmebg2TMB5t9yg7juzggyUfFMbTKJaSuye6VTkbV66ESAFuyWHdNcBvhRiLMcacsXwncqq6CWgFDMO54GENTr+4z4C2qrotEAEWF6WiwhnZpx11q5Sh/+hUFmw6/Ux0TJlYLr71Tm5+/n9UrFGLn4a9xpinBrNj/Vqv+2xTrQ1dErswcvFIth22X39BVK4ZS4M2VVk41apyJcQTwFUiMhm4Hecf0u4iMhq4FngqmMEZY4yvfBl+5N8448e9oKrnqWojVT1HVR9X1ZxOVRgP5UtH8mG/9lQqE8Wto2azeof306dVE+txw79eoOvAB9m3bSsfPfYAP48cxrHDp7d/sO2DZGomr819LdDhF1vJ3RM5eSyD+T9bVa64U9UZOEMnRQNv4lzs8C+cO9NcoqopQQzPGGN85sup1SuAH4G9IpIqIq+KyJUiUjVAsRVL1crF8FG/swkPE3qPmMWWfUe9tpOwMJpd0Inb/vcOrS7tzoKJExj14ACW/Pozqn9dNFwrthZ9mvXhu7XfsWjnosJ6GsVK5Vqx1G9TlYW/bObYYavKFXeqOlNVOwLlgHigrKp2UNWZQQ7NGGN85sup1RZAFeBq4BfgXGAssFVElovIu4EJsfhJrFKGD25rz6Fj6fQeMYs9h3Me/iKmTCydbhvATc+9Svmq1fhx6KuMffpRdm5Y92ebfi36UaVUFV5IeeGUJM/kX7seTlVuwc92BWtJoapHVTVNVe1qIWNMkeXTPaJUda+qfqOqg4C/4VTppgONcG59U+Tt37+fkycDX5VpVrM87/VJZsveo/QdNZtDx9NzbV+tbn16PTOESwfcx54tmxn92P388v5wjh85TJnIMtzX+j4W7FzAj+t/DHjsxZFTlYtj4ZRNVpUzxhhTZPjSR66ciHQTkWdFZDqwH/gAZ/y4QTgVuiItMzOTsWPHMmzYMDZv3hzw451drzJv3tiGxWkHuHN0KsfTcx/5QMLCaHHRpfT93zu07NSVuT9+y8gH7mTp9F/oWb8nTSo14ZU5r3As/VjAYy+O2vWoywmryhljjClCfKnI7QE+x7li9SMgWVXjVPUKVX1ZVWcHIsDCFBYWxiWXXEJ6ejojRoxg8uTJpKfnXik7U52bVuOFq1syc/VuHvh0PhmZeZ8aLRVblktuH8jNz75Kubiq/PDmy3z2r38ysNYtbDu8zYYjKaDKtWKp39qqcsYYY4oOXxK5FCAC6AxcCnQWkdYiIrlvVrTUq1ePu+66i6SkJGbMmMHw4cPZunVr3huegWvaxvNEjyb8sHgbT3y1ON/93KrVa8CN/36Jzv3vZfeWTcwdMpxrNrfig3kj2XFkR0BjLq6Ss6pyU6wqZ4wxJvT5crHDuUAF4DpgIXAZTv+4vSIyQUQeDUyIhS8mJobLL7+cG2+8kSNHjvDuu+8ydepUMjICN+j77R3rMfDC+oyZvZGXJq7I93YSFkbLTl247dVhtLj4UmIX7afblEq8/enTduFDAVSJj6Ve6zgW/mxVOWOMMaHP14sdjqrqL6r6L5x7Ft4EzAe6As8GIL6gatSoEQMHDqRZs2ZMnTqV9957jx07AlfpGtSlMb3a1+atX9bw3nTvgwDnpFTZcnS+4x5u+s/LlK5YgdifNvL+/93Prk0bAhRt8dWuRyInjmWw0Kpyxkci0lhE5ntMB0TkgWxtLhSR/R5tngxWvMaYos+Xix2qi8i1IvK6iMzDuV/hF0BF4C2gV4BiDKrSpUtz9dVXc91117F//37eeecdZs6cSWYOt846EyLCf65oQddm1fnP98v4Yo7vF1xUb9CIO154hwVJx9ixYR2jH72PXz8ayYmjNsJCflWJL0u9pDgWTNnM8SNWlTP5p6orVDVJVZOAtsAR4EsvTadntVPVZwo3SmNMceJLRS4N5x6rycBEnKFHqqhqK1W9V1XHBSLAUNG0aVMGDhxIo0aNmDRpEqNGjWL3bv/f0CI8THitVxIdGlRm8BcLmbx0u8/7KBdTjsuvvovP/raRcm0ak/rteEY9OIDlv02z0635lNwjkRNH01kwJfBXL5tiqxOwRlWtLG6MCRhfErlOQAX39lyPqur3qnr6DUOLsdjYWK677jquuuoqdu7cydtvv82sWbP8Xp2Ljgjnnd7JNKtZjrs/mcustb4njFc0uILEag0Zl7iAq//1LKUrVOT7117k8/88we4tdsowL3G1y1K3VRUW/LzJqnKmoG4AxuSw7lwRWSAiP4hIM28NRKS/exed1J07dwYuSmNMkebLxQ6/qGrABygTka4iskJEVovIY17W3yoiOz36l9we6JiyHZ+WLVsycOBAEhMT+eGHH/jwww/Zu3evX48TGx3BqFvbUatiKW7/IJUlaft92j48LJzB7QaTdjiNSSdmcdOzr9Cp30C2r1vNh4PuZdon73PimPfbgxlHux51OXE0nYW/WFXO+EZEooCewGdeVs8FElS1FfAG8JW3fajqcFVNVtXkuLi4wAVrjCnSfLrYIdBEJBynv103oCnQS0Saemk61qN/yXuFGqSrXLly3HTTTfTs2ZO0tDTefvtt5syZ49dTl5Vjoxnd72xiYyLoMzKF9bsO+7R9+xrt6VSnE+8uepfdx/aQdGl3bvvfcJr+7SJSvv6cUQ/dxco/Ztjp1hzE1fGoyh0N7HiCptjpBsxV1dP6RqjqAVU95D6eAESKSJXCDtAYUzyEVCIHtAdWq+paVT0BfApcHuSYciQitGnThoEDB1KrVi2+/fZbPv74Yw4cOOC3Y9SqUIrR/dqTkZlJ75Gz2HHAt6Low20f5mTmSV6f9zoApcuVp8uA++n17yGUKluOb199ni+efZI9aVZ18qZdj7ocP5JuV7AaX/Uih9Oq7oVj4j5uj/M57P8Ot8aYEiHUErlagOc35mZ3WXZXi8hCEflcRGp721Fh9i+pUKECvXv3pnv37mzYsIGhQ4eyYMECv1W6GlQty6i+7dl96AS3jJzNfh/6bNUuV5veTXrz9eqvWbp76Z/LazZqws3PvcrFtw1g2+qVfPDIPUwf8wEnj9ntvTzF1SlLYkurypn8E5HSOAOnj/dYNkBEBriz1wCLRWQB8Dpwg1pZ3BhTQKGWyHm7S0T2D7hvgURVbQlMxrnf6+kbFXL/krCwMNq3b8+AAQOIi4vjyy+/ZOzYsRw6dMgv+0+qXYHhvZNZs/MQt32QwtET+R+c+I6Wd1AxpiIvprx4SnIZFhZO6y6X0ffVYTQ5/wJmf/UZox66i1WzfrPTrR7aX+ZU5Rb9YlU5kzdVPaKqlVV1v8eyYao6zH38pqo2c6/4P0dVfwtetMaYoi7UErnNgGeFLR5n2JM/qepuVT3uzr6LM1ZTyKhcuTJ9+/alc+fOrFq1iqFDh7JkyRK/7Pv8hlV47YbWzN24l7s+nsPJjPxdLVs2qix3J93NnO1zmLxx8mnry1SoSNeBD3L9v14gpkwZvnnlWcY/9xR7t27xS9xFXVZVbv7kTZywqpwxxpgQEmqJXArQUETquld93QB849lARGp4zPYElhVifPkSFhZGhw4duPPOO6lQoQKfffYZn3/+OUeOnPmgvN1b1OA/VzRn6oqdDPpsAZmZ+aucXdXwKhpWbMjLqS9zPOO41zbxZzXj5udf46Jb+5O2cjkfPHI3M8eO5uRxO93arkei01fOrmA1xhgTQkIqkVPVdOAe4CecBG2cqi4RkWdEpKfb7D4RWeL2L7kPuDU40eatatWq9OvXj4suuoilS5cydOhQVqzI/31Uc3LT2QkM6tKYr+an8cx3S/N1GjQiLILB7Qaz5dAWPlr6UY7twsLDadOtJ7f97x0an9uRP8aP5f2HB7I65Y8Sfbq1akI5EltUZv7kjVaVM8YYEzKkJHw5Jycna2pqalBj2Lp1K1999RXbt28nKSmJrl27EhMTU+D9qSr//m4ZI2eu46HOjbivU8N8bXfvlHtJ2ZbCd1d+R5VSeY94sHnpYn4e+Ta7Nm2gbutkLr71TipUr5HndsXRjg0H+Oy5VM6+vB7J3RKDHY7Jg4jMUdXkYMfhD6HwGWaMKTy+fH6FVEWuOKtRowZ33HEHHTt2ZMGCBQwdOpQ1a9YUeH8iwhM9mnBV61q8Mmklo//I312AHkl+hOMZx3lz3pv5ah/ftDk3P/8aF95yB1uWL+H9RwYyc9zHnDzh/fRscVY1oRwJVpUzxhgTQiyRK0QRERF06tSJfv36ERUVxejRo/nuu+84frxgSVFYmPDCNS3pdFZVnvx6Md8uSMtzm4RyCdx41o2MXzWe5XuW5+s44RERtO1xOX1ffYeG7c/jjy/G8MHDA1kzZ3aB4i7K2vWoy/HD6Sycan3ljDHGBJ8lckEQHx/PnXfeybnnnktqairDhg1jw4aC3Vc7MjyMt25qQ3JCRR4aN59pK/MeM+/OVndSPro8LzNKdQAAACAASURBVMx+wad+b7EVK9HjvkFc9+SzRERF89WLz/Dli8+wb/u2AsVeFFVLLEdCc7cqd8yqcsYYY4LLErkgiYyMpEuXLvTt2xeAUaNG8eOPP3LypO83aI+JDOe9Pu2oHxfLnaPnMHdj7vd9LRdVjnuS7iF1eypTNk7x+Xi1m7Wk9wuvc8HNt7FpySLef/gufvngXfaklYzhSrKqcousKmeMMSbI7GKHEHD8+HEmT55MSkoKlStX5sorryQ+Pt7n/ew4eIxrh/3O/qMnGXfnuTSqVjbHtumZ6Vz77bUcSz/G11d8TVR4VIFiP7hnFzM++YDlv00jMyODhJatadW5G/Xbnk1YeHiB9lkUfPvGAnasP0Dv/55LVExEsMMxXtjFDsaYosoudihioqOj6dGjB7179+bkyZOMGDGCyZMnk57u26m7qmVjGH3b2USGh3HLiNls3pvzuHX/396dh0lRnXsc/77ds7Pv2zAsCiIoIIIBFMEFVDRu0YgxblERtxsTrzcm8Qpi9JIYNRqDLFFxi8aoLAoKKBGXuLCjgCggy8jAwDAMy6w9c+4f3cBMM1szM73M/D7PM09VV52qevtYHt4+VacqzhPHvYPuJf1AOq+se+WYY2/SsjUX3HkPYyfP4PSrrmXPD+nMeewRpt/5Cz5781UOZO855n1Hs0EXdSX/YJF65UREJKLUIxdl8vPzee+991i5ciVt27blsssuo0OH0B73sS5jHz+d+hltGify+rghtG6cWGHZOz/wX2Kt7uNIqlJSXMym5UtYtXAem1ctx+P1cvygIfQfNZrU3icTeFd4vfD2X1eSuXm/euWilHrkRCRWqUcuhiUlJXHppZdy9dVXk5uby/Tp01m8eDHFxdV/t+qJHZry3A2D2J6Txw3Pf8n+/Irvu7tn4D0U+Ar428q/1Ub4gcRtMD/53UR+8ZepnHLBxWz9aiWvT/wdM+65nRXvvU1B7sFaOVakDbqwG/kHi/h6ccO4N1BERKKPErkodcIJJ3D77bfTu3dv/v3vf/P3v/+dzMzMam8/qGtLJl8zgHUZ+7nlxaXkF5WfCHZr1o0xvcbw1ndvsX5Pzd86UVqLDp0Yce1NjJ3yAufddjcJSUksen4qU8ddz8LpT5O5eVOtHi/c2ndvRlrvlqxYqBGsIiISGbq0GgPWrl17+HlzZ599NkOGDMHjqV4OPnNFOr/65ypG9W7H5GsGEOc9erucghwunHkhvVr2YvrI6XV6+XPHxu9YtXAe33yyGF9RIR17nkj/UaPpMfgM4uLj6+y4dWXHphze/NMyhlx2HAPO6xLpcKQUXVoVkVilS6v1TO/evbn99tvp0aMHCxcu5PnnnycrK6ta2152SioPXNSbBWt38ruZX5X73Lhmic24vd/tfJHxBR9u+7CWoy+r/XE9OG/cLxk75QVGXHczeftzmPf0Y0y77Xo+/scMcjJ31unxa1v77s3oHOiVKyqo/uVvERGR2qAeuRjinGP16tW8++67+Hw+Ro4cyaBBg6rVO/fYgvX8ddEGbh3end9ecOJR64tKirhizhX4SnzMumQW8d7w9I65khK2fL2KVQvmsXHpFzgc3U8ZSL9Ro+nabwAeT/Q/wiRjYw5vPbqMIZcfx4BR6pWLFuqRE5FYFUr7paF2McTM6NevH926dWPOnDm8++67rFu3jksuuYQWLVpUuu2vR/Zkz8FCpi7eRMuUBG4dflyZ9fGeeO4ddC+3vX8b//jmH1zf5/q6/CqHmcdD176n0LXvKezbvYuvFs3nqw/mM3PSgzRr246+517ASWeNJKVps7DEcyw6HOfvlVu5cCsnD08lPjH6k08REakf1CMXo5xzLF++nPnz5wNw3nnnMWDAgErvbysucfzXayuYuzqDP/2kLz8d1PmoMre9fxurMlfxzuXv0DKpZZ3FX5liXxEblnzOqgXz2Lb2K7xxcfQcMoz+o0bToUevqHyEyaFeuaGXH88po9IiHY6gHjkRiV2htF9K5GJcdnY2s2fPZvPmzRx//PFcfPHFNG3atMLyBb5ibn5hKZ9u2M0zPz+V8/q0L7N+095NXD7ncq7oeQX3D76/rsOv0u5tW1i18F3WfvQBhXl5tOnanf6jRnPi6SOIT0qKdHhlzHlyBbvTD3DtH4aqVy4KKJETkVilwQ4NSIsWLbjuuuu44IIL2Lx5M5MnT2bVqlXlDmoASIzzMuXnp3JyanPuenUFn20sO2iie/PuXHXCVfzr23/xXfZ34fgKlWrduQvn/GIct055kXNvvgNKSlg47WmmjLuORc9PJSt9W6RDPGzQhd3I21/E1x/puXIiIhIe6pGrR7Kyspg1axbbtm2jV69eXHTRRTRu3LjcstkHC7ly6mfsyMnntbGDOanTkXvQ9ubv5cKZF9K7VW+mjZwWVZcynXNs//YbVi2Yy7eff0Kxz0fnPn3pP2o0xw0cjDcusrd9zv7LCrJ+OMC1Dw8lPkG9cpGkHjkRiVXqkWugWrVqxY033sjIkSP57rvvmDx5MmvWrCm3bItGCbx002k0S47n+ue+ZNOuA4fXNU9qzu39b+fzjM/5KP2jcIVfLWZGpxNOZPRd/83YyTM44+rrycncwdtPTGL6nb/g09dfYX/W7ojFN+gif6/cGvXKiYhIGKhHrp7KzMxk5syZZGRkcNJJJzF69GhSUlKOKrdx1wGunPIZyfFe3rxtKO2b+e87Kyop4vLZlwPw1sVvhe1xJMeipKSYzSuXs3LBXL5fuQwz4/iBg+k3cjRpJ/cLe4+ieuWiQ6R65MxsM7AfKAZ8wTGY/4R8EhgN5AI3OOeWV7bPhtiGiTRk6pET2rZty80338xZZ53F2rVrmTx5MuvXH/0KruPaNOaFG09jb24h1z77BXtzC4EjjyPZvG8zr61/Ldzhh8Tj8dJ9wCAuv28CNz05nYEXXca2dV/zxsP38/yvxrFs7mzyDxyoeke15NC9cuqVa9DOcs71r6AhvgDoEfgbCzwT1shEpF5RIlePeb1ehg8fzi233EJKSgqvvvoqs2bNIj8/v0y5k1ObMf26gWzJyuXGGUvILfS/N3RYp2EM7TiUZ1Y9Q3Z+diS+Qsiat2vPmdfcyK2TZ3DBHb8mqUkTPnxxOlNvu575U55i56YNdR5Dxx7N6XRCC5Yv2EpRod72IEe5BHjR+X0ONDezDpEOSkRikxK5BqBDhw6MHTuWYcOGsWrVKiZPnszGjRvLlBl6fGueuro/q7btZdzLyyn0lWBm3DvwXg4WHWTyyskRiv7YxCUk0PvMs/nZQ3/m55Oe5MRhI/jmP4t5+bd388rvf82axR9QVFhQZ8c/7aKu5O0rZO3H2+vsGBK1HLDAzJaZ2dhy1ncCSg+3Tg8sK8PMxprZUjNbumvXrjoKVURinRK5BiIuLo5zzjmHm266ifj4eF566SXmzp1LQcGRZOb8kzrwyGUn89G3u/j16yspLnEc3+J4rux5Jf/69l9s3LuxkiNEr3bdjmPU2Lu49ZkXOOuGsRTm5vLe5CeYdtsNLH75OfbuyKj1Y3bs0YJOJzRn+fwt+NQr19Cc7pwbgP8S6h1mdmbQ+vJu2jzqZmXn3DTn3EDn3MA2bdrURZwiUg8okQtS3wd/pKamMm7cOAYPHsySJUuYMmUKW7ZsObx+zGlp/Ob8XryzOoMJc9bgnOOO/neQEpfCo0sfjWDkNZfUqDEDLriYGx5/hiv/9xHS+vRl2dxZPPvLW3jzkQfYsPQLSkpqL+kadGE3cvcVska9cg2Kc257YJoJzAROCyqSDpR+rUoqoJNERI6J3rUaZOWqG8nN3UxKSleSk7uQktKVlOQuJCd3ITm5Mx5PQqRDrLH4+HjOP/98evXqxaxZs3j++ecZMmQIZ599NvHx8Ywb3p3s3EKmfbSJlo0S+NXInozrN45Hlz7Kx+kfMyx1WKS/Qo2YGWkn9SXtpL4c2JPF6g/m89UH7zH70Ydo0roN/QLvd23UvPL311alU88WdOrp75XrM6wjcRrBWu+ZWSPA45zbH5gfBUwMKjYHuNPMXgN+BOQ452q/W1hEGgQ9fiTI1m3Psy9nJbl5W8jL24zPt7/UWg9JSR1JSe5KckqXUtNDSV5i3XyBOlRQUMDChQtZunQprVu35tJLLyU1NRXnHPe+sZo3lqUz4ce9uWZwKpfNuQyPeXjz4jeJ90Tv40iORbHPx6ZlX7JywVy2fr0KjzeOHj8aSv9Ro+nUq88xP8Lkh/XZzHpiBWf8tAf9zj763bZSdyLx+BEz646/Fw78P5T/4Zx72MzGATjnpgQeP/I0cD7+x4/c6JyrtIHS40dEGha9azXIsTaCzjmKirLJy9tCbu5m/zRvM3m5/qnPt69UaSuT5CUnH0n0kpPS8HqjO8nbuHEjs2fPZv/+/ZxxxhkMHz4czMO4l5fz/rqdPDmmP81afcddi+7ivtPu45oTr4l0yHUm64dtrF74LmsWf0BB7kFad+5Cv1EX0nvYCBKSj34WX1VmPracvZm5XPuHIcTFq1cuXPRmBxGJVUrkgtRVI1hUlE1ubtnkzp/0bcHn21uqpJGU2CGoF69r4HJtGl5vdLz8PT8/n/fee4+VK1fSrl07Lr30Ulq0bsv1z33Jsi3ZTLvuVF7dej9rs9Yy7/J5NEtsVvVOY1hRfj7f/OcjVs6fS+bmjcQnJdN72Fn0GzWaNmldq72f9PXZzH5iBcOu6kHfs9QrFy5K5EQkVimRCxKJRrCoaK//8mzu5sD0SKJXVFT6mWxGYmJ7/+XZQ/fjHU700vB6k8MaN8D69et5++23yc3NZfjw4fQbNJifTf+STbsPMGlMa8Yvu5mre13NfafdF/bYIsE5x44N37JywVzWf/YxxUVFdOrVh/6jRtPjR0PxxlV+mdk5x6zHV5CTmcvP1SsXNkrkRCRWKZELEm2NYFFRzuHLtYfuxcvN3RJI8vaUKZuY2D5wmbZLYABG18C0bpO83Nxc5s2bx9dff03Hjh0ZPupCbv3Xt+w5WMi5Z37CovS3eevit+jevHudxRCNcvflsObD91n1/rvk7NxBSrPmnHz2efQ99zyatm5b4Xbp3+xh9l9WMuyqnvQ9KzWMETdcSuREJFYpkQsSSiNYmHEQShyeJC+W6MWTFIfFhe8pLUVF+/yJXZnePP/0qCQvoV2pXrzS0zS83tDv5SrPmjVreOeddygsLOTUIcN48IsiSjwH8aRN4tR2A5h8bmw9KLi2uJISNq9ewcoFc9m0fAmG0f3UQfQfOZoufU/BPGXPGeccMx9bzr5deeqVCxMlciISq0Jpv/T4kSB7Z35H4db9ZRd6LZDYxeFJ9GJJXjyJcYGpF0vyLz88f6hs6WQw0YvFe6oc/Rgf35T4+L40bdr3qHU+3/4jgy5KDb7YtfsDioqyypRNSGh75BEqpRO95C7ExTWqdn306dOHLl268M477/DlJx9ybfuO/GNHO3zZI/nYN4tPfviEMzqdUe391Rfm8dCt/6l0638qOZk7Wf3Be3y1aAEbl35B83Yd6DvyAk4acS7JTZr6y5sx6KJuzPnLStZ9msHJI9QrJyIiNaceuSCF2/ZTvK+QkgIfrqCYkvxiXIEvMC2mJD+w/NB8fjElBT7wVaMePRxOBitK9konhmUSxlLLLN6LecomhD7f/rL34uVuOXzZtrBwd5myCQltyj4jr0yS17jc0J1zrF69mnnz5uErLubLwo6kd3mZ1JaNmXnJW8R59JvAV1TEd1/+h1UL5vLDN2vxxsfTa+iZ9Bs1mvbH9QT8I1j37c7n2oeG4I3X87jrknrkRCRWxfSlVTM7H3gS8AJ/d85NClqfCLwInApkAVc55zZXts9wNILOV0JJQTEu3xeYFleaDJYUBCWGgaTQFZVUfTADS6h+MujifRSxmwIyKShJJ69kK3m+7zno+47Coswyu05IaF3m0Smln5UXF9eEffv2MWfOHDZs2MAO87Ci01zuGPJfXNen/j6O5Fjs2vI9qxbOY+3HH1KUn0e77sfTb+RoGrc6mXnPrOPMMT3VK1fHlMiJSKyK2UTOzLzAt8BI/K+xWQJc7ZxbW6rM7UBf59w4MxsDXOacu6qy/cZSI+iKnT/pKwhK+g4lhvnFZRPGMslg2eXVYQkeSHC4+GJcXD7Fcbn4PPsosj34PNmUePMoicujOC4fS/SS0Kgp8Y1bsi8P1m7YQXaBhy9StvK3G/5Oq0Y1exNCfVSQm8u6j//NygVzyUrfSmKjRiQ2OhlPfF9u+ONF6pWrQ0rkRCRWxfI9cqcBG5xzmwACr7C5BFhbqswlwITA/BvA02ZmrpYy0peuOoVWWQVVF4x6hseTiNebhMebHJgm4fUm4/EEpt6ko9d7/MsTvd1I9p6I15uM2dHJRmOgY2D+yizIe+hr0sP6/WJHa5pybvzV0K3s8oz//TQyAcWwJbvmctmzk6ouKCLSQERbItcJ2Fbqczr+dxGWW8Y55zOzHKAVUOZGMDMbC4wFSEtLq6t4o5ijpCSfkpJ8KNpbdfFKmCcBb6kk7+gksPxkT8pXjIfcpLaUmEauhqqgcGukQxARiSrRlsiVN6QzuKetOmVwzk0DpoH/skR1A7j2nyuqW7R+y8uGzZ/ApsXw/WLYvQ4Al9iS7F0nsGvRBpzz0PqOO2h14w1YfP1696qIiEgsiLZELh0o/Q6jVGB7BWXSzSwOaAbsQWqmKA+2fnYkcctYBa4E4htBl6Ew4DryCjuz468vk79mDY2GDaP9+AdISNUN+yIiIpESbYncEqCHmXUDfgDGAD8LKjMHuB74DLgCWFRb98c1KMU+2L4Cvv/Qn7xt+xKKC8ATB6mDYPhvoNtw6HQqxflF7HrqSbJf/j3eVi3p9MTjNDn//CqfiSciIiJ1K6oSucA9b3cC8/E/fuQ559waM5sILHXOzQGeBV4ysw34e+LGRC7iGOIcZK7z97ZtWgxbPoWCff517U+G026B7iMgbQgkHnmW3P7332fHQ3/Al5lJi6vH0Obuu/E2bRqRryAiIiJlRVUiB+CcmwfMC1r2QKn5fODKcMcVk/ZuPXKpdNNiOBh4ZlyLbnDST6D7cOh6JjRqddSmRdu3s+MPD3Ng0SISTziB1KeeJLlfvzB/AREREalM1CVyUgMHs/xJ26HELft7//JGbf1JW7fh/mnzikfxOp+PPS+9zK6//hWco+2999Lyums1mEFERCQKKZGLZQUHAgMUPvQnbju/8i9PbApdTocfjfMnbm16QTXuZ8tbvZqM8RMoWLeOxiNG0P5/7ye+U6e6/Q4iIiJyzJTIxRJfIfyw9Mjl0vQlUOIDbwJ0/hGcfT90GwEdTwFv9f/TFu/fz64n/kL2q68S16YNnZ56kiYjR2owg4iISJRTIhfNSkpg59f+HrfvF8OWz6DoIGDQsT8Mvct/uTRtMMQnh7x75xz7589n58OP4Nu9mxbXXEObu3+Jt3HjqjcWERGRiFMiF02cgz2bjtzjtvljyM3yr2vdE/r/LDBA4QxIrtl7TQvTf2DHQxM5uPgjknr3JnXyZJJPPqkWvoSIiIiEixK5SNu/80ji9v1iyAm8oaxJR+hxXmCQwpnQtGPl+6kmV1TEnhdeYNfTf8M8Htr99j5aXHMNFqdTQUREJNboX+9wy8+BzZ8eSd52+V99RVJz6DYMTv+l/3lurY6v1gCFUOSuWMGO8RMo+PZbGp97Du1//3viO3So1WOIiIhI+CiRq2tF+bDtiyOJ2/bl/ldfxSX7723rN8bf69a+L3jq5iXqxTk5ZD7+BHtff5249u1J/dvTNDnnnDo5loiIiISPErnaVlIMGSv9SdumD/1JnC8fzAudToVh9/gHKHQ+DeIS6zQU5xz75s5j56RJFO/ZQ8vrr6fNXXfiadSoTo8r0lCZWWfgRaA9UAJMc849GVRmBDAbCDzokbeccxPDGaeI1B9K5GrKOdj97ZF73DZ/7L98CtC2Nwz8hT9x6zIUksL3aqvCrVvZ8eBEDn76KUknnUTnqVNI7tMnbMcXaaB8wD3OueVm1gRYZmYLnXNrg8p97Jy7KALxiUg9o0TuWOSkH0ncvv8I9mf4lzdPgxMv9t/j1u1MaNw27KG5wkKynnue3c88g8XF0e7++2lx9RjMWzeXbUXkCOdcBpARmN9vZuuATkBwIiciUiuUyFVH7h5/T9uh5C1rg395Sit/wtZ9hL/XrWW3SEZJ7rJlZIwfT+GGjTQ57zza/e63xLdrF9GYRBoqM+sKnAJ8Uc7qIWa2CtgO/Ldzbk05248FxgKkpVX8Wj0RadiUyJWnMBe2/udI4paxGnAQ3wi6ng6n3ugfoNC2D3g8kY6W4r17yXzsMfb+6w3iO3YkdcozNBkxItJhiTRYZtYYeBO42zm3L2j1cqCLc+6AmY0GZgE9gvfhnJsGTAMYOHCgq+OQRSRGKZEL9q8bYN07UFIEnnhIHQQj7vP3unU6FbzR8/J45xz73n6bnZP+SHFODi1v+gVt7rgDT0pKpEMTabDMLB5/EveKc+6t4PWlEzvn3Dwzm2xmrZ1zu8MZp4jUD0rkgrXuCYPH+d9Z2mUIJETnCM+C779nx8SJ5H72OUn9+pL23LMk9eoV6bBEGjTzv6D4WWCdc+7xCsq0B3Y655yZnQZ4gKwwhiki9YgSuWBn/S7SEVSqpLCQrOnTyZo6DUtMpP2E8TT/6U+xKLjEKyKcDlwLfGVmKwPLfgekATjnpgBXALeZmQ/IA8Y453TpVESOiRK5GHLwyy/ZMX4Chd9/T9PRo2n32/uIa9Mm0mGJSIBz7hOg0leyOOeeBp4OT0QiUt8pkYsBvuxsMv/0KDkzZxKfmkrn6dNpPOyMSIclIiIiEaZELoo558iZOYvMP/2J4gMHaDV2LK1vG4cnOTnSoYmIiEgUUCIXpQo2bWLH+AnkLllC8oABtJ8wnqSePSMdloiIiEQRJXJRpqSggKypU9k9/e94kpNpP/FBml9xhQYziIiIyFGUyEWRg599xo4JD1K4ZQtNf/xj2v3mf4hr3TrSYYmIiEiUUiIXBXxZWez84x/ZN+dt4rukkfbcszQaOjTSYYmIiEiUUyIXQa6khL1vvknmnx+jJDeX1rffRqtbb8WTmBjp0ERERCQGKJGLkILvviNjwoPkLVtGysCBtH9wAonHHRfpsERERCSGKJELs5L8fHY/M4WsZ5/F26gRHR5+mGaXX4b/zT4iIiIi1adELowOfPwJOyZOpGjbNppdeilt/+de4lq2jHRYIiIiEqOUyIWBb9cudv7fJPbNm0dC166kzZhBo8E/inRYIiIiEuOUyNUhV1LC3tdfJ/Oxx3H5+bS+605a3XILnoSESIcmIiIi9YASuTqSv/5bdjzwAHmrVpEyeDDtxz9AYrdukQ5LRERE6hElcrWsJDeX3ZMnk/X8DLxNm9Lxj5NoevHFGswgIiIitU6JXC06sHgxOx6cSNH27TS74ie0vece4lq0iHRYIiIiUk8pkasFRTsz2fnII+yfP5+E446jy8svkTJwYKTDEhERkXouahI5M2sJ/BPoCmwGfuqcyy6nXDHwVeDjVufcxeGKMZgrLib71dfY9cQTOJ+PNnffTatf3IhpMIOIiIiEQdQkcsB9wAfOuUlmdl/g82/KKZfnnOsf3tCOlr92LRnjJ5D/1Vc0Ov102o9/gIS0tEiHJSIiIg1INCVylwAjAvMvAB9SfiIXUSUHD7Lrr0+z58UX8bZoQcc//5mmF47WYAYREREJu2hK5No55zIAnHMZZta2gnJJZrYU8AGTnHOzyitkZmOBsQBptdRTtn/RInY89Ad8GRk0v+oq2v76V3ibNauVfYuIiIiEKqyJnJm9D7QvZ9XvQ9hNmnNuu5l1BxaZ2VfOuY3BhZxz04BpAAMHDnTHFHBAUUYGOx5+mAPvf0Bijx50+sc/SBlwSk12KSIiIlJjYU3knHPnVrTOzHaaWYdAb1wHILOCfWwPTDeZ2YfAKcBRiVytxOvzkf3KK+x68ilcSQlt7vk1rW64AYuPr4vDiYiIiITEE+kASpkDXB+Yvx6YHVzAzFqYWWJgvjVwOrC2LoLJ++prNv/0Knb+3ySSB55K93fepvUttyiJE5FKmdn5ZrbezDYEBm4Fr080s38G1n9hZl3DH6WI1BfRdI/cJOB1M7sJ2ApcCWBmA4FxzrmbgROBqWZWgj8JneScq9VErvjAAXY9+RTZr7yCt1VLOv3lCZqcd54GM4hIlczMC/wNGAmkA0vMbE5QO3UTkO2cO97MxgB/BK6qjePvOLiDTTmbamNXta9GN7iEiR2a+GcOtfuHP2NH/VtQXtng7Q6XLV2mgmNgZY9XbpkKYjtqefD3KVUmeFl5372ifQd/d6meRG8izZOa1/p+oyaRc85lAeeUs3wpcHNg/j/AyXUYA1t+fi0F69fT4uoxtPnVr/A2aVJXhxOR+uc0YINzbhOAmb2Gf0R+6UTuEmBCYP4N4GkzM+dcjVOdj9I/4qHPH6rpbkSkDozsMpLHRzxe6/uNmkQuGpgZbf7rLuJatSK5X79IhyMisacTsK3U53TgRxWVcc75zCwHaAXsLl3oWEben9X5LHq06HFMgYdDNPfguECX4aF8uqLP5ZYpVfZwucOTSsqUs7+K9l16fxXGWqpsZWWCfzOUV7Y6ZSQ0nRp3qpP9KpEL0uTssyMdgojErvIyleB/8apT5phG3rdJaUOblDbVKSoi9UQ0DXYQEYl16UDnUp9Tge0VlTGzOKAZsCcs0YlIvaNETkSk9iwBephZNzNLAMbgH5FfWukR+lcAi2rj/jgRaZh0aVVEpJYE7nm7E5gPeIHnnHNrzGwisNQ5Nwd4FnjJzDbg74kbE7mIRSTWKZETEalFzrl5wLygZQ+Ums8n8HglEZGa0qVVERERkRilRE5EREQkRimRExEREYlRSuREREREYpQSOREREZEYpUROREREJEZZQ3gOwi2yNAAABtdJREFUZeBdht8FLW4G5JQzD9CaoPce1qLgY9XmdlWVqWh9ecuDl1X1WXUW+3VW3W0qKxfqurqssy7OuZh/X5WZ/Rh4DthSanFl9RRt51Uo2x3LuVWd86q8ZfWhzuqy/Spvmeqs8mWRab+cc/X+D5hW2bLg9fgf3Bm2WGpru6rKVLS+qvqp5mfVWYzXWXW3qaxcqOuiqc6i9S/Ueou28yqU7Y7l3KpO/dTXOqvL9kt1VrP6qeBzndRZQ7m0+nYVy8pbX1eO9VjV2a6qMhWtr6p+qvO5LqnOQncsx6ruNpWVC3VdNNVZtDqWeqsrdfn/YlXlavL/YnnL6kOd1WX7Vd4y1VnlyyJSXw3i0mqozGypc25gpOOIJaqz0KnOQqc6q5rqKHSqs9CpzkJXV3XWUHrkQjUt0gHEINVZ6FRnoVOdVU11FDrVWehUZ6GrkzpTj5yIiIhIjFKPnIiIiEiMUiInIiIiEqOUyImIiIjEKCVyIiIiIjFKiVyIzKy7mT1rZm9EOpZoZWaNzOwFM5tuZtdEOp5YoPMqdGZ2aeAcm21moyIdTyzQeVY9asNCp3MrdLXVhjWoRM7MnjOzTDP7Omj5+Wa23sw2mNl9le3DObfJOXdT3UYafUKsu8uBN5xztwAXhz3YKBFKnTXU8ypYiHU2K3CO3QBcFYFww0rtV82oDQud2rDQRaINa1CJHDADOL/0AjPzAn8DLgB6A1ebWW8zO9nM3gn6axv+kKPGDKpZd0AqsC1QrDiMMUabGVS/zsRvBqHX2f2B9fXdDNR+1cQM1IaFagZqw0I1gzC3YXHHumEscs59ZGZdgxafBmxwzm0CMLPXgEucc/8HXBTeCKNXKHUHpONvCFfS8H4sHBZina0Nb3TRKZQ6M7N1wCTgXefc8rAGGgFqv2pGbVjo1IaFLhJtWIM9QUvpxJFfXuD/H7hTRYXNrJWZTQFOMbPf1nVwUa6iunsL+ImZPYPelRms3DrTeVWpis6zu4BzgSvMbFwkAosCar9qRm1Y6NSGha5O27AG1SNXAStnWYWvu3DOZQEN9R+NYOXWnXPuIHBjuIOJERXVmc6rilVUZ08BT4U7mCij9qtm1IaFTm1Y6Oq0DVOPnD8z7lzqcyqwPUKxxBrVXehUZ6FTnVVMdVMzqr/Qqc5CV6d1pkQOlgA9zKybmSUAY4A5EY4pVqjuQqc6C53qrGKqm5pR/YVOdRa6Oq2zBpXImdmrwGfACWaWbmY3Oed8wJ3AfGAd8Lpzbk0k44xGqrvQqc5CpzqrmOqmZlR/oVOdhS4SdWbOVXg7hYiIiIhEsQbVIyciIiJSnyiRExEREYlRSuREREREYpQSOREREZEYpUROREREJEYpkRMRERGJUUrkJOaY2Q1m5syscYjbTTCz3XUVl4hIdagNk9qkRE5EREQkRimRExEREYlRSuQk6pjZEDObY2bbzeygma00s2sqKd81cJniZ2b2kpntN7NMMxtfQflTzOxzM8s1sxVmNixo/XVm9omZ7TGzbDP7t5kNrO3vKSL1k9owCSclchKNugCfAjcDPwbeBJ43s6ur2O5RIBe4ApgOjDezO4LKpAAvAFOBnwAFwEwzSylVpivwInAl8DMgHfjIzLrX4DuJSMOhNkzCJi7SAYgEc869dmjezAz4CEgFbgFerWTTNc65WwPz882sLfA7M3vGOVcSWJ4M3O2cWxTYfwawAjgTeC9w/Imlju8BFgKDgJ8Dh9eJiJRHbZiEk3rkJOqYWQsze8rMtgBFgb+xQM8qNp0Z9PktoCP+BvSQIuDDUp/XBqaHy5jZiWY208x2AsWBbU6oxvFFRNSGSVipR06i0QxgMPAQ/kZqH3AbcEkV22VW8LkDsDUwv6/UL1ucc4X+H8wkAZhZE2ABsBP4NbAFyAf+fqiMiEgVZqA2TMJEiZxEFTNLAi4E7nTOTSm1vDq9x20r+JwRQghD8P+yHemc+6bU8ZuFsA8RaaDUhkm46dKqRJtEwIv/Bl7g8C/Mi6ux7WVBny/H3wCmh3D85MC09PGH4r95WESkKmrDJKzUIydRxTmXY2ZLgAfMbB9QAtwH5ABNq9i8j5lNxT9C7EzgJuCXpS9DVMPnwAFgupn9Cf8v2wnADyF9ERFpkNSGSbipR06i0c+A7/EPn38Sf6P2YjW2+x/8DeWbwK347095OpQDO+d24h+y3x6YDdwNjAM2hLIfEWnQ1IZJ2JhzLtIxiNSImXXF32j+2Dn3TmSjEREJjdowqQn1yImIiIjEKCVyIiIiIjFKl1ZFREREYpR65ERERERilBI5ERERkRilRE5EREQkRimRExEREYlRSuREREREYtT/A0N93CmmKW7XAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, (ax1, ax2) = plt.subplots(1, 2, sharey=False, figsize=(10,5))\n",
    "\n",
    "ax1.plot(alphas, coefs_l)\n",
    "ax1.set_xscale('log')\n",
    "ax1.set_xlabel('alpha', fontsize=15)\n",
    "ax1.set_ylabel('weights', fontsize=15)\n",
    "ax1.set_title('Lasso coefficients as a function of the regularization')\n",
    "ax1.axis('tight')\n",
    "\n",
    "ax2.plot(alphas, errors_l)\n",
    "ax2.set_xscale('log')\n",
    "ax2.set_xlabel('alpha', fontsize=15)\n",
    "ax2.set_ylabel('error', fontsize=15)\n",
    "plt.axis('tight');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Similar to Ridge : Lasso is not using our features, it is therefore not relevant to use here. But it is striking to see that None of our feature is needed for the model, which feels wrong"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[12.073253405708334, 19.739279902600973, 0.011777967717193749],\n",
       " [12.115436289653879, 19.734037124717766, 0.012302844064883689],\n",
       " [12.274891982322856, 19.869035689518462, -0.0012568520643245762],\n",
       " [12.276023039960382, 19.869783371373966, -0.0013322090849863244]]"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "errors_l"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 7) Time cross validation with Ridge Regression"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ridge does not do a better job that a simple Linear Regression but we will try to improve its performances with a cross validation over our time series"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Perform grid search for alpha coefficient over splits of crossed validation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looking for the influence of cross-validation settings over the model performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Getting back the objects:\n",
    "with open(path + '/data/objs_cv_ridge.pkl', 'rb') as f :\n",
    "    splits, rmse_split, mae_split = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cross validation with 2 time splits\n",
      "\n",
      "Model regression for split number 1 : RMSE : 22.52, MAE : 12.82, R² : 0.02\n",
      "\n",
      "Model regression for split number 2 : RMSE : 19.51, MAE : 12.02, R² : 0.01\n",
      "Mean RMSE 21.02\n",
      "Cross validation with 3 time splits\n",
      "\n",
      "Model regression for split number 1 : RMSE : 20.31, MAE : 11.77, R² : 0.02\n",
      "\n",
      "Model regression for split number 2 : RMSE : 22.11, MAE : 12.77, R² : 0.03\n",
      "\n",
      "Model regression for split number 3 : RMSE : 19.83, MAE : 12.07, R² : 0.01\n",
      "Mean RMSE 20.75\n",
      "Cross validation with 4 time splits\n",
      "\n",
      "Model regression for split number 1 : RMSE : 19.41, MAE : 11.20, R² : 0.02\n",
      "\n",
      "Model regression for split number 2 : RMSE : 23.34, MAE : 13.19, R² : 0.02\n",
      "\n",
      "Model regression for split number 3 : RMSE : 19.78, MAE : 11.97, R² : 0.02\n",
      "\n",
      "Model regression for split number 4 : RMSE : 20.44, MAE : 12.28, R² : 0.01\n",
      "Mean RMSE 20.74\n",
      "Cross validation with 5 time splits\n",
      "\n",
      "Model regression for split number 1 : RMSE : 19.82, MAE : 11.27, R² : 0.02\n",
      "\n",
      "Model regression for split number 2 : RMSE : 21.05, MAE : 12.08, R² : 0.02\n",
      "\n",
      "Model regression for split number 3 : RMSE : 23.77, MAE : 13.64, R² : 0.02\n",
      "\n",
      "Model regression for split number 4 : RMSE : 18.25, MAE : 11.42, R² : 0.00\n",
      "\n",
      "Model regression for split number 5 : RMSE : 20.66, MAE : 12.39, R² : 0.01\n",
      "Mean RMSE 20.71\n",
      "Cross validation with 6 time splits\n",
      "\n",
      "Model regression for split number 1 : RMSE : 19.60, MAE : 11.29, R² : 0.02\n",
      "\n",
      "Model regression for split number 2 : RMSE : 19.57, MAE : 11.35, R² : 0.02\n",
      "\n",
      "Model regression for split number 3 : RMSE : 23.74, MAE : 13.46, R² : 0.02\n",
      "\n",
      "Model regression for split number 4 : RMSE : 21.15, MAE : 12.54, R² : 0.03\n",
      "\n",
      "Model regression for split number 5 : RMSE : 18.46, MAE : 11.40, R² : 0.01\n",
      "\n",
      "Model regression for split number 6 : RMSE : 21.20, MAE : 12.63, R² : 0.01\n",
      "Mean RMSE 20.62\n",
      "Cross validation with 7 time splits\n",
      "\n",
      "Model regression for split number 1 : RMSE : 19.96, MAE : 11.58, R² : 0.02\n",
      "\n",
      "Model regression for split number 2 : RMSE : 18.37, MAE : 10.87, R² : 0.02\n",
      "\n",
      "Model regression for split number 3 : RMSE : 22.03, MAE : 12.49, R² : 0.02\n",
      "\n",
      "Model regression for split number 4 : RMSE : 24.37, MAE : 13.96, R² : 0.02\n",
      "\n",
      "Model regression for split number 5 : RMSE : 19.62, MAE : 11.97, R² : 0.02\n",
      "\n",
      "Model regression for split number 6 : RMSE : 17.76, MAE : 11.12, R² : -0.00\n",
      "\n",
      "Model regression for split number 7 : RMSE : 21.71, MAE : 12.86, R² : 0.01\n",
      "Mean RMSE 20.55\n",
      "Cross validation with 8 time splits\n",
      "\n",
      "Model regression for split number 1 : RMSE : 19.90, MAE : 11.31, R² : 0.02\n",
      "\n",
      "Model regression for split number 2 : RMSE : 19.57, MAE : 11.26, R² : 0.02\n",
      "\n",
      "Model regression for split number 3 : RMSE : 20.42, MAE : 11.75, R² : 0.02\n",
      "\n",
      "Model regression for split number 4 : RMSE : 23.16, MAE : 13.23, R² : 0.02\n",
      "\n",
      "Model regression for split number 5 : RMSE : 23.57, MAE : 13.69, R² : 0.03\n",
      "\n",
      "Model regression for split number 6 : RMSE : 17.98, MAE : 11.31, R² : -0.00\n",
      "\n",
      "Model regression for split number 7 : RMSE : 18.09, MAE : 11.19, R² : -0.00\n",
      "\n",
      "Model regression for split number 8 : RMSE : 22.13, MAE : 13.07, R² : 0.01\n",
      "Mean RMSE 20.60\n",
      "Cross validation with 9 time splits\n",
      "\n",
      "Model regression for split number 1 : RMSE : 19.71, MAE : 11.27, R² : 0.02\n",
      "\n",
      "Model regression for split number 2 : RMSE : 19.27, MAE : 11.11, R² : 0.02\n",
      "\n",
      "Model regression for split number 3 : RMSE : 19.48, MAE : 11.29, R² : 0.02\n",
      "\n",
      "Model regression for split number 4 : RMSE : 22.27, MAE : 12.63, R² : 0.02\n",
      "\n",
      "Model regression for split number 5 : RMSE : 24.21, MAE : 13.88, R² : 0.02\n",
      "\n",
      "Model regression for split number 6 : RMSE : 21.24, MAE : 12.63, R² : 0.03\n",
      "\n",
      "Model regression for split number 7 : RMSE : 18.20, MAE : 11.32, R² : 0.01\n",
      "\n",
      "Model regression for split number 8 : RMSE : 18.03, MAE : 11.15, R² : -0.00\n",
      "\n",
      "Model regression for split number 9 : RMSE : 22.60, MAE : 13.31, R² : 0.01\n",
      "Mean RMSE 20.56\n",
      "Cross validation with 10 time splits\n",
      "\n",
      "Model regression for split number 1 : RMSE : 19.42, MAE : 11.17, R² : 0.02\n",
      "\n",
      "Model regression for split number 2 : RMSE : 20.26, MAE : 11.53, R² : 0.02\n",
      "\n",
      "Model regression for split number 3 : RMSE : 18.44, MAE : 10.88, R² : 0.02\n",
      "\n",
      "Model regression for split number 4 : RMSE : 21.15, MAE : 12.02, R² : 0.02\n",
      "\n",
      "Model regression for split number 5 : RMSE : 23.08, MAE : 13.21, R² : 0.02\n",
      "\n",
      "Model regression for split number 6 : RMSE : 25.23, MAE : 14.52, R² : 0.02\n",
      "\n",
      "Model regression for split number 7 : RMSE : 18.28, MAE : 11.49, R² : 0.00\n",
      "\n",
      "Model regression for split number 8 : RMSE : 18.25, MAE : 11.30, R² : 0.01\n",
      "\n",
      "Model regression for split number 9 : RMSE : 17.78, MAE : 11.05, R² : -0.01\n",
      "\n",
      "Model regression for split number 10 : RMSE : 23.09, MAE : 13.56, R² : 0.01\n",
      "Mean RMSE 20.50\n"
     ]
    }
   ],
   "source": [
    "splits = [2, 3, 4, 5, 6, 7, 8, 9, 10]\n",
    "mae_split = []\n",
    "rmse_split = []\n",
    "r2_split = []\n",
    "\n",
    "# Model\n",
    "model = linear_model.LinearRegression()\n",
    "model = linear_model.Ridge()\n",
    "\n",
    "for split in splits :\n",
    "    print(\"Cross validation with %i time splits\" % split)\n",
    "    errors_splits, alphas_splits = time_cross_validation(model, X, y, 5, split, gs=False)\n",
    "    \n",
    "    errors_array = np.array(errors_splits)\n",
    "    \n",
    "    mae_split.append(errors_array[:,0].mean())\n",
    "    rmse_split.append(errors_array[:,1].mean())\n",
    "    r2_split.append(errors_array[:,2].mean())\n",
    "    \n",
    "    print(\"Mean RMSE %.2f\" % errors_array[:,1].mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqMAAAFACAYAAACIp3XZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xm81nP6x/HXVadFQtZRWi1jC1HIMipZElGWMJaSsZ6xjSV7iy1lGUbDkAihLA2yTFGUSQtJUtlLmRTRL0nL6Vy/Pz73qbtz7nOf7306517OeT8fj/vRub/rdYbHuLo+n8/1MXdHRERERCQTamQ6ABERERGpvpSMioiIiEjGKBkVERERkYxRMioiIiIiGaNkVEREREQyRsmoiIiIiGSMklERERERyRgloyIiIiKSMUpGRURERCRj8jIdQGWqUaOGb7bZZpkOQ0RERKRMK1eudHevdoXCKp2MbrbZZvz222+ZDkNERESkTGb2e6ZjyIRql32LiIiISPZQMioiIiIiGaNkVEREREQyRsmoiIiIiGSMklERERERyRgloyIiIiKSMUpGRURERCRjlIyW18CBvH3TeJo3hxo1oHlzePum8TBwYKYjExEREckZSkbL6e3/O5D97uxOi/njcYcW88ez353defv/Dsx0aCIiIiI5w9w90zFUms0339wrawem5s1h1/lv8xIn8wLdOYlX6M5Ivm3WgXnzKuWVIiIiUoWZ2Up33zzTcaSbktFyqlEDavlqfmYbNmcl/bmFPvTHDAoLK+WVIiIiUoVV12RUw/Tl1LQpHMokDKcQyGcw7RlP06aZjkxEREQkdygZLachZ41nJN05j6E4NXiLToykO0POGp/p0ERERERyhpLRcjpqq2l8cuNIpjQ7g9fowtGM5Y0/P8NRW03LdGgiIiIiOUNzRivAqtfGUvfEY7hrr6e54bOzK/19IiIiUvVozmgamFkTMxtvZnPM7DMzuyJ2/LTY90Iza5Pk/k5m9rmZfWVm16cv8uTqHt+Rn7f/I+1nD2bixExHIyIiIpI70j1MXwBc7e57Am2BfDPbC5gFnAxMKO1GM6sJDAaOA/YCzozdm3k1alD/unwOYTLD//ZRpqMRERERyRlpTUbdfZG7T4/9/CswB9jJ3ee4++dl3H4Q8JW7f+Pua4DngZMqN+Loal/QgzW1N+egDwczodSUWkRERETiZWwBk5k1B/YHpkS8ZSdgQdz3hbFjxZ97oZl9aGYfFhQUbGqY0W21FTXOOZszeY57b1yavveKiIiI5LCMJKNmVh94CbjS3ZdHvS3BsRKrr9z9UXdv4+5t8vLyNiXMlOVdkc9mrGK3/z7Bu++m9dUiIiIiOSntyaiZ1SIkosPd/eUUbl0INIn73hj4X0XGtsn22Yd1h/2Jv9Z8mH59tA2TiIiISFnSvZregMeBOe5+X4q3TwN2M7MWZlYbOAN4taJj3FQ1L8un+bpv2GzCW6qOioiISIUys6FmtsTMZsUdG2Rmc81sppmNMrMGSe6vaWYfm9nouGMtzGyKmX1pZiNieVbapLsyehhwDnCkmc2IfTqbWTczWwgcArxuZv8BMLNGZvYGgLsXAH8F/kNY+DTS3T9Lc/xl69YN37EhV9cZTJ8+UIXbuIqIiEj6PQl0KnZsLNDS3fcFvgBuSHL/FYQ8Kt7dwP3uvhvwC3B+xYQajZreV4a+ffH+/dnVv+Sxd3bhyCPTH4KIiIjklqhN72OLwEe7e8sE57oBp7r7WQnONQaGAXcAf3P3E2Kj1j8CO7p7gZkdAvR192M37beJTtuBVoYLL4SaNbm2/sP07avqqIiIiKRNL+DNUs79HbgOiF/Ysi2wLDYCDaV0K6pMSkYrQ6NGWLdu9CwcyocTVzJuXKYDEhERkRyQV9SeMva5MJWbzewmwgZDwxOcOwFY4u7Fd+eJ1K2oMikZrSz5+dRd+QuXNHhe1VERERGJoqCoPWXs82jUG82sB3ACcJYnnoN5GHCimc0jbBx0pJk9A/wENDCzon6Yae9WpGS0shxxBOy9N723GMz77zvvvJPpgERERKQqMrNOQG/gRHdfmegad7/B3Ru7e3NCR6Jx7n52LHEdD5wau7QH8Eoawl5PyWhlMYO//pUdFkyny/ZTVB0VERGRTWZmzwEfALub2UIzOx94CNgCGBvrVPRI7Nr1XYnK0Bv4m5l9RZhD+nglhZ+QVtNXphUrYKed+GL3Luw+7RnGjIGjj85cOCIiIpK9oq6mr2pUGa1M9etDjx7s9skLtGq0RH1HRURERIrZpGTUjFI7/EvMpZdia9bwSJshfPABjB2b6YBEREREskekYXozLgG2cGdg7HsrYDTQEJgBnOTOwsoMtDwyPkxf5Kij8C++YGf/hh0b5zFpUphSKiIiIlJEw/TJXQYsj/v+IGHZ/1mxZwyo4Liqlvx8bMECHjlhNJMnw5gxmQ5IREREJDtErYyuALq4M96M7YEfgI7uvGvGycBD7jSq5FhTljWV0YIC2HlnCnfbnRZfjaVhQ/jgA1VHRUREZANVRpNbDdSO/dwBWAlMjH3/GTR3NKm8PLj4YmqMe5uB589lyhR4661MByUiIiKSeVGT0alAvhl7A5cDb7mzLnZuZ9LcqT8n/eUvULs2py7+J82aob6jIiIikpvM8jCrU+zYMZhdidkBqT4uajJ6DbAX8CnQBLgp7tzpwH9TfXG1s8MOcNpp1HxmGH2vWcHUqfDmm5kOSkRERCRlI4CH138zuxx4C7gLmIzZCak8LFIy6s5n7uwKbA80d+eLuNPXAFen8tJqKz8fli/nHHuG5s1VHRUREZGc1BaI39npWuBe3DcDhrBx0bJMkZJRM4aa0cKdpe4UT5+WQ2j5JGVo2xb235+ajwzm5pucadPgjSibdImIiIhkj20Ji9nBbB+gEfBI7NwLhNH0yKIO0/ckVEUT2Q7okcpLqy2zUB2dNYseO0+kRQtVR0VERCTnLAaax37uBMzH/evY982AwlQelsoOTKWlTC2BH1N5abV25pmw9dbk/WswN98MH34Ir7+e6aBEREREInsBuBuzQUBv4Km4c/sDX6bysFL7jJpxBXBF7GszQjl2dbHL6gJ/AJ505/xUXpwOWdNntLhrroEHHmDtV/PZ48hGbL01TJumvqMiIiLVWc70GTXLA24EDiTsxHkb7mti514GJuF+T+THJUlGjwaOAQz4G/AssKjYZWuAucBI9xKJasZlbTL69dew225w66080awvvXrBq69Cly6ZDkxEREQyJWeS0QoWdQemPsAQd76v/JAqTtYmowCdO8OMGRR8PZ899qnFVluFIXtVR0VERKqnnElGzdYBh+A+NcG51sBU3GtGfVzU1k79ci0RzXr5+bBoEXmvjeLmm2H6dHjttUwHJSIiIlKmZKWzWkBBSg9LMkw/ErjBna9jPyfj7pyeyovTIasro+vWhaH6Jk0oeOc99tgDttwSPvpI1VEREZHqKKsro2ZN2bCC/l3gEmBOsavqEjostcZ9j6iPzktybntCdguwA6WvppfyqFkTLr0Urr2WvDmfcsst+9CzJ7zyCnTtmungRERERDZyHtCHkA868Tswbex34C+pPDjSnNFcldWVUYCff4addoIePSh46BH22gs23zwM2as6KiIiUr1keWV0e0Jx0oCZwFmxP+OtAb7DPaVF7UpGM61XLxg5Er7/nqdf3Ypzz4WXX4Zu3TIdmIiIiKRTViej8cyaAYvWt3Pa1MclmTN6aSoPcuefFRFQRcqJZPSjj6BNG3jgAQouvZy99oJ69UJ1tEYqWxKIiIhITsvqZNSsHu4r1/9clqJrozw6STKaylZO7k7kJfzpkhPJKIQ965ctgzlzeGa4cc458NJLcPLJmQ5MRERE0iXLk9EN7ZzMCilrLVEKrZ00TJ8Nnn4azj0Xxo6loP1R7L031KkDM2aoOioiIlJdZHky2gMYjftSzHpSdjI6LPKjlYxmgVWroGlTOOwwGDWK4cPh7LPhxRfhlFMyHZyIiIikQ1Yno5UocjJqRm2gJ3AQ0JCwNegUYJg7FTKBtaLlTDIKcOONcPfd8O23rNupKXvvDbVqwSefqDoqIiJSHSgZTXaRsSfwFtAI+AhYQljefwDwA9DJndmVGGe55FQy+t130KIF9O4Nd97Js8/CWWfBCy/AqadmOjgRERGpbFmdjJpNI5We8+4HRX50xGR0IrAVcII738Udbwq8DvzizhGRA0yTnEpGIXS7nzQJFixgXV4dWraEvDxVR0VERKqDKMmomQ0FTgCWuHvL2LFBQBdCn8+vgfPcfVmx++oCE4A6hE2PXnT3PrFzTwLtgP+LXd7T3WcUe/GTpJaMnhf10qjJ6O/Ame78O8G5bsCz7mwWOcA0yblkdOxYOOaYsKDp7LN57jn4859DG9LTTst0cCIiIlKZIiajRwArgKfiktFjgHHuXmBmdwO4e+9i9xmwubuvMLNawPvAFe4+OZaMjnb3Fyv+typb1HrbPMJ+o4nUhQ3VUtkEHTvCH/8IgwcD0L077Lkn9OsHhak02hIREZEqyd0nAD8XOzbG3QtiXycDjRPc5+6+Iva1Vuyz6avYzQyz7bHy7x0ZNRm9HrjdjIM3fj9tgf5A74R3SWpq1Aj71U+eDB99RM2acOut8NlnYe6oiIiIVGl5ZvZh3OfCcjyjF/BmohNmVtPMZhDW/ox19ylxp+8ws5lmdr+Z1SnzLWadMZsErCKsH1qF2STMjk814KjD9NOAZsC2sV+gaAHTDsBSQuV0PXciT1qtTDk3TA+h+X3jxqEsOnQo69bBvvuGUzNnQs2s21pAREREKkLUBUxm1pwwrN6y2PGbgDbAyZ4kwTOzBsAo4DJ3n2VmDQkJZW3gUeBrd++fJICLgH8C7wAvsyEvPBnoCFyK+7/K+j3WPy5iMvpE1AcCuBN50mplyslkFODii2HYMFi4ELbdlpEj4fTT4bnn4IwzMh2ciIiIVIZNSUYtNKW/GOjoEbbiNLM+wG/ufk+x4+2Ba9z9hCQ3zwfewP2SBOceATrj3rSsGNbfks6m92bWBHgK2BEoBB519wfMbBtgBNCcUGXt7u6/JLj/bqCo/Hubu49I9r6cTUY//TSUQwcNgmuuobAwfC0sDKdUHRUREal6ypuMmlkn4D6gnbv/WMo92wNr3X2ZmW0GjAHudvfRZtbQ3RfFFjndD6xy9+uTBLAC6Ib72ATnjgZG4V6/rN+jSLkbBpmxtRmtzCh7XsEGBcDV7r4n0BbIN7O9CHNS33H33Qgl3xL/A1iYg3AA0Ao4GLjWzLYsb/xZbZ994E9/gocfhsJCatSAPn1gzpywsl5ERESqJzN7DvgA2N3MFprZ+cBDwBbAWDObYaE6iZk1MrM3Yrc2BMab2UxgGmHO6OjYueFm9inwKbAdcHsZYYwntIJKpB2hhVT03yniMH0/oI57SBLNOBJ4BahHmGNwjDufpfLi8Bx7hfA/4ENA+1hW3hB41913L3bttSEGvz32/XHgP+5eanqWs5VRgBEjwpj8669D584UFsJ++0FBAcyapeqoiIhIVZPVTe/jhernEOAN4N9smDPaDTgO+Avw/frr3ZNujBQ1Gf0KuN2dJ2PfPyYkof2AO4Df3Dkxtd/DmhMy55bAd+7eIO7cL+6+dbHrjwH6AEcTkuCpwGB3v7e0d+R0MrpmDTRrBgccEBJSwl71p50Gw4eH/qMiIiJSdeRQMlq84aQDVuw7sWOOe9ISWl7E1zYCvgnvpwmwH3CRO1PNuA9SW+BkZvWBl4Ar3X15lNZU7j7GzA4EJgE/EkrUBcWvi7VBuBCgdu3aqYSVXWrXhosugv794euvYZddOPnkMILfv39Y0KTqqIiIiGRAh4p8WNTK6GLgL+68ZkYP4D53to2d6wi84k6kiaqxrv+jCUPs98WOfU4Zw/QJnvMs8Iy7v1HaNTldGQX43/9CdfSKK+CesNjtpZfCXvXPPBP2rhcREZGqIWcqoxUs6gKm94DrzTgeuIYwX7TIH4EFUR4SW6X1ODCnKBGNeRXoEfu5R7HnF91b08xiCbDtC+xLWAlWdTVqBN26wdChsDJ0aejWLays798/zB8VERERSSuzHTBrEffdMLsQs79j1iXVx0VNRq8CVgPPA8uAm+LOnUv0VVOHAecAR8ZWe80ws87AAOBoM/uSMCd0AICZtTGzIbF7awETzWw2oSHr2XFbX1Vd+fnwyy/w/PMA61fWf/HF+kMiIiIi6fQkITcs0o/QBL8TMAqznqk8bJP7jJqxJbDKnTWb9KBKkPPD9ADuYaJonTrw4YdgRmFhWNe0ciXMng15UWf+ioiISNbKmWF6sx+Ai3H/N2Y1gMXAINwHYtYPOAn3VlEfV+4+o0XcWZ6NiWiVYRaqo9Onw5SwhWxRdfTLL+HZZzMcn4iIiFQ3WxG2gwdoDWwDDI99HwfsmsrDNjkZlTQ45xzYckt46KH1h7p2hVat4LbbNHdURERE0mohsFfs5+OBubgX9RXdCliVysOUjOaC+vWhRw944QVYsgQIBdM+feCrr0LfUREREZE0GQoMxOwF4DrCWp4ibYE5qTwsrXvTp1uVmDNaZO5c2HNPuOMOuPFGIEwnbd0ali8PpzV3VEREJHflzJxRALNzgQOBGcBQihLKsBXpB7gPi/woJaM55KijwjL6b75Zn3m++iqcdBI88QT07JnZ8ERERKT8cioZrUApDdObcZwZt5jxqBlNY8eOMKNR5YQnG8nPhwULYPTo9Ye6dAkr62+7DdauzWBsIiIiIuUQKRk14w9mTAFeIzSlPx/YLnb6POCWyglPNtKlCzRpAoMHrz9kBn37hmLpM89kLjQRERGR8ohaGf0HUB/YI/aJ30z+baBjBcclieTlhf3q3347TBKNOeGEMHdU1VERERHJNVGT0U7Aze58BRSfZLoQ2KlCo5LSXXAB1K4N//zn+kNF1dFvv4Wnn85caCIiIiKpSmXO6LpSjm8H/F4BsUgUO+wAp50Gw4bBihXrDx9/PLRpA7ffruqoiIiI5I6oyehE4DIzasYdK6qQ9iJ025d0yc8P/ZziJonGV0efeipzoYmIiEg1YLYvZiMw+xqz1ZgdEDt+B2bHpfSoKK2dzGgJvA8sAkYBvYF/AS1jn7bufJHab1H5qlxrpyJFDUbXroWZM0MmGjt88MHw44/w+edhNF9ERERyQ860dgrJ5qvAJEJBsg/QBvfpmN0KtMW9c9THRaqMujOLsPfoh0BPwpD9ycAC4OBsTESrtKL96mfNgokTNzrcty/MmxdG8UVEREQqwV3Ak7i3A+4odm4G0CqVh6npfa5auRIaN4ajj4YRI9Yfdoe2bWHx4tAfX9VRERGR3JBDldFVwAm4v41ZTWAtGyqj7YG3cK8b9XHamz5X1asHvXrByy/D//63/nBRdXT+fHjyyYxFJyIiIlXXEmDnUs7tDXyXysMiV0bN6A50I7RxKpHtunNQKi9OhypdGQX4+mvYbTe49daQgca4wyGHwKJF8OWXqo6KiIjkghyqjA4EzgVOBT4gVEZbA78R+s8/jnu/yI+LuIBpAHAdMA34ClhT/Bp3zov60nSp8skoQOfOMGNGKIXWqrX+8H/+A506wSOPhD75IiIikt1yKBmtA7wEHAf8ADQk9J3fERgDdMM9cqPJqMnoEuB+d+4qT8yZUi2S0ddfD1swjRgB3buvP+wOhx4K338fqqN16mQwRhERESlTziSjRcw6Enbh3A74GXgH97EpPyZiMvo9cJ47Y1J9QSZVi2R03bowVN+kCbz33kanxoyBY4+Fhx+Giy/OUHwiIiISSc4loxUkajJ6HdAGON29xHagWataJKMAgwbBddeFnqP77LP+sDscdhgsWABffaXqqIiISDbLuWTUrC7QiARriXCfHfkxKSxgugfoArwHLCvxSqd31JemS7VJRpcuDW2eevQIk0TjjB0LxxwTtrK/5JIMxSciIiJlipKMmtlQ4ARgibu3jB0bRMjR1gBfA+e5+7Ji99UFJgB1gDzgRXfvEzvXAnge2AaYDpzj7iXWB8U9rDHwKHBsorOA414zwbnEj4tYGT0LGAYUAj9ScgGTu5e6xD9jqk0yCqHN08iRYZLoVlutP+wOhx8O332n6qiIiEg2i5iMHgGsAJ6KS0aPAca5e4GZ3Q3g7r2L3WfA5u6+wsxqEXbWvMLdJ5vZSOBld3/ezB4BPnH3h5MEMZ7Q2mkApSxsx/29EsdKEbXP6ABgBLCtOzu506LYJ+sS0WonPx9++63E1ktm0K8fLFwIQ4ZkKDYRERGpEO4+gbBYKP7YGHcviH2dDDROcJ+7+4rY11qxj8eS1COBF2PnhgFdywijDXAZ7g/jPhb390p8UhA1Gd0SGOrOr6k8XNKodeuwMf0//xnKoXE6dgzV0bvuglWrMhSfiIiIlCXPzD6M+1xYjmf0At5MdMLMaprZDELT+rHuPgXYFlgWl8wuJPSUT2Y2UK8csSUUNRl9CehQUS+VSpKfD59/Du+8s9Hhol2Zvv9e1VEREZEsVuDubeI+j6Zys5ndBBQAwxOdd/d17t6KUDk9yMxaEuZ4lri0jFddBvTG7LBU4itN1DmjpxOG6t8BxlFyARPuvFERAVWkajVnFELZs0mTUAYdNWqjU+7Qrl3YtOnrr6Fu5B1jRUREJB2irqY3s+bA6KI5o7FjPYCLgY7uvjLCM/oQdky6l7AeaMfYnNNDgL7unmhxUtHNtYF/AH8hzBctOXLuvkNZMRTJi3jdc7E/e8U+JV4JRF41JZWkbl244AK4++6wYqlp0/WniqqjHTvCY4/BZZdlLkwRERGpOGbWCegNtCstETWz7YG17r7MzDYDjgLudne3sCDpVMKK+h7AK2W8cghwGmGeaeIFTKnEH7Ey2qysa9yZvymBVIZqVxmFkIS2aAG9e8Odd250yh3atw87Mn3zjaqjIiIi2STiavrngPaEXY8WA32AGwgtm5bGLpvs7hebWSNgiLt3NrN9CYuTahKmaY509/6xZ+7MhtZOHwNnu/vqJEH8ClyL+yOlXpOCyH1Gc1G1TEYBunaFSZNCt/tivZzefRc6dIAHHoDLL89MeCIiIlJSzjS9N/sG+CvuFTJFs9QFTGYbVkmZUa+sT0UEIxUkPx9+/BFeeKHEqfbtw9zRu+6C339Pf2giIiKS8/oD12BWvyIeVmpl1Ix1wCHuTDWjkDJWVrln35zRalsZLSyEPfeEbbaBDz4ocbqoOvr3v8MVV6Q/PBERESkphyqjLwBtgc2BD0mwMyfup0d+XJJktAcw2p2lZvSk7GR0WLLzmVBtk1EI4/BXXgkffhh6kBbToQPMnRvmjm62WQbiExERkY3kUDI6vsxr3CO3BNWc0apq2TLYaSc4/XQYOrTE6ffeC0P2998fclYRERHJrJxJRitYpKb3Znxjxn6lnGtpxjcVG5ZssgYN4Jxz4LnnYOnSEqfbtQvV0QEDYGWZ3chEREREKkfUHZiaE1oGJFKPBHugShbIzw+N8J94IuHpfv1g8WJ4pEIaM4iIiEiVZXYpoVdp0c/JP6k8Osmc0S2BBrGv84CuwIxil9UldPvv5k6LVF6cDtV6mL7IEUeEfUC//BJqlPy7x1FHwaefwrffQj31RBAREcmYrB6mNysE2uI+NfZzMo575IXtySqjVxGS0G8Ji5dGxX6O/8wBrgQejPpCSbP8/LBK6a23Ep7u2xeWLIGHH05vWCIiIpJD3GvgPjXu52SflDosJauM7gb8ETDgVeAa4PNil60BPnfnuxR/pbRQZRRYswaaNYMDDoDXX094ydFHw8yZIWfdPDv/PiYiIlLlZXVlNJ7ZEcB03FckOLc50Br3CZEfF3E70HbAR+6UfGkWUzIa06cP3HZbGKrfZZcSp//7Xzj8cBg0CK65JgPxiYiISC4lo+uAQ9ZXSjc+1xqYWlHD9Ou5815FJKJm1sTMxpvZHDP7zMyuiB3fxszGmtmXsT+3LuX+gbH75pjZg2ZmmxpTtXDRRVCzZqlj8YcdFqqjAweCcncREREpQ7L8qz6QUp+etPYZNbOGQEN3n25mWwAfERZG9QR+dvcBZnY9sLW79y5276HAIOCI2KH3gRvc/d3S3qfKaJzu3eHtt2HhwoQrlSZNCknpwIFw7bUZiE9ERKSay+rKaBiabx/71hcYAiwsdlVd4HjgN9wPjfrovAoILzJ3XwQsiv38q5nNAXYCTmLDLzgMeBfoXfx2wi9Zm5CR1wIWV3rQVUV+ftir/vnnoVevEqcPPRSOOSYko5dcAvUrZLdZERERqSIOBi6L/ezAaUBBsWvWAHOBlMpaGduBycyaAxOAlsB37t4g7twv7l5iqN7M7gH+QkhGH3L3mxJccyFwIUDt2rVbr169ulLizznusM8+UKdO2CI0wQyHDz4ISendd8N112UgRhERkWosqyuj8cy+Bbri/klFPK7MOaNm1DXjMTPaVsQLwzOtPvAScKW7L494z67AnoQG+zsBR1ooGW/E3R919zbu3iYvL62F3+xmFqqj06fDlCkJLznkEDj22LCQaUVOLVUTERGRtHFvUVGJKERIRt1ZBZxBGCLfZGZWi5CIDnf3l2OHF8fmkxbNK12S4NZuwGR3X+GhlcCbUHEJcrVw9tmwxRYweHCpl/TtCz/9lPQSERERkQoTdTvQcUCHTX1ZbPX748Acd78v7tSrQI/Yzz2AVxLc/h3QzszyYgltO0LTfYlqiy2gZ08YOTJ0uk+gbVvo1ClUR3/9Nb3hiYiISPUTNRkdDJxnxj1mHGnG3mbsFf+J+JzDgHMIQ+wzYp/OwADgaDP7Ejg69h0za2NmQ2L3vgh8DXwKfAJ84u6vRXyvFLn00tAIf8iQUi/p1w+WLoWHHkpjXCIiIlItRW16X3wP0vibDHB3Utr6KR3U2qkURx0FX3wRtlwqZV7t8cfD5Mkwb14oqIqIiEjlypkFTBUsamW0Q7HPkXGfou+SK/LzYcECGD261Ev69IGff4Z//CONcYmIiEh2M6uL2WrMulbYIzPV2ikdVBktRUEB7Lwz7L47jB1b6mUnnBDaPX37LWy5ZRrjExERqYZypjJqtgC4BPfSq1opiFoZjb2b48yIFmiDAAAgAElEQVS4xYxHzWgaO3aEGY0qIhhJk7y8sEXo22/D3LmlXqbqqIiIiCTwL+BywoLyTRZ1zugfCCveWwPzgBbAge5MN+MJYJU7l1REQBVJldEkFi+GJk3g4ovhwQdLvaxLF/jvf8PcUVVHRUREKk8OVUbvAf5MWEP0DmFHzPiE0im2rXsyUSuj/yBsfL9H7BO/fc/bQMeoL5Qs8Yc/hP3qhw1L2uG+Tx/45Zek+aqIiIhUL6cAqwnbf/4JOJWwPWj8J7KoyWgn4GZ3vmLjzBdgIWFHJMk1+fmwfDk880ypl7RpE6qj994L//d/aYxNREREslPYgSnZZ+dUHpfKnNF1pRzfDvg9lZdKlmjbFvbfP2y3lGS6Rt++sGyZqqMiIiKZZmZDzWyJmc2KOzbIzOaa2UwzG2VmDRLc18TMxpvZHDP7zMyuiDvX18y+L9YDPm2iJqMTgcvMNuolWpS99CLs0CS5pmi/+lmzYOLEUi874AA48US4776QlIqIiEjGPEkYsY43Fmjp7vsCXwA3JLivALja3fckbKeeb2bxmxbd7+6tYp83yozCbF/MRmD2dazV0wGx43dgdlwqv1DUZLQ3cCAwC7iNkIheYMYE4BDg5lReKlnkzDNh663L3Ixe1VEREZHMc/cJwM/Fjo1x94LY18lA4wT3LXL36bGffyVsqV6+aZYh2fwI2BF4CohfVb8auCyVx0VKRt2ZBbQBPgR6EobsTwYWAAe780UqL5UsUq8enHcevPwy/O9/pV62//5w0kmqjoqIiGS5XsCbyS4ws+bA/sCUuMN/jQ3zDzWzrct4x13Ak7i3A+4odm4G0CqVgCPPGXXnK3fOcaeRO7Xd2dGds9z5MpUXSha65BJYtw4efTTpZX36hEVMf/97muISERGpXvLM7MO4z4Wp3GxmNxGG44cnuaY+8BJwpbsvjx1+GNiFkEQuAu4t41V7ACNiPxdfdLIc2CaVuCMlo2b0M2PPVB4sOWTXXaFTp5CMrl1b6mX77w9du4ZkVNVRERGRClfg7m3iPsmrRHHMrAdwAnCWl9JE3kKT+peA4e7+ctFxd1/s7uvcvRB4DDiojNctAUpbMb838F3UuCF6ZfQiYJYZn5pxoxm7pPISyQH5+bBoEYwalfSyvn1DdfT++9MTloiIiCRnZp0I63tOdPeVpVxjwOPAHHe/r9i5hnFfuxHWCCXzPNAfs8Pjjjlmf4zFUWplNmFsEXdgqgG0B06PBbkt8DHwHPCCe2oZcLpoB6YUrFsHu+0WdmV6772kl55ySthJdN68sPZJRERENl2UHZjM7DlCTrYdYeejPoTV83WApbHLJrv7xWbWCBji7p0tJI4TgU+Bwth1N7r7G2b2NGGI3gk7bV7k7ouSBFGHUGE9DvgBaEjoO78jMAbohnvpQ63FHxclGd34/dQEjiIkpl2BrYAP3Dk86Y0ZoGQ0RYMGwXXXwcyZsM8+pV42cybstx/ccgv075/G+ERERKqwnNkOtIhZR8IunNsRVvi/g/vYlB+TajK64f3UIayovwfY0X2jHqRZQcloipYuhcaNoUcPeOSRpJcedBB8+GH4uWlTuOMOOOusNMQoIiJSReVcMlpBUtmBCTNqmdHFjGcIk1eHAbOBlFZ7SZbadtvQd/SZZ5Lu/Tl8OHz6adi0yR3mz4cLLwzHRUREpIozm4fZ3ZjtXyGPizhntBMbhuW3BN4nLOl/wZ0fKyKQyqDKaDl89FHYkP6BB+DyyxNe0rx5SECLq1ED9t4btt8+fLbbbuM/43/edluoVavkM0RERKqrnKmMmt0DnAo0Ab4mLGgaiXtZC58SPy5iMloITC16mTuld0fPIkpGy6lt29C7ac6csGVoMTVqlL6V/Yknwo8/wk8/hT+TtYBq0CBxolrasfr1E4YjIiJSJeRMMlrE7BBCsfJUwiKmOYRi5fO4R+5DHzUZbe7OvPJFmjlKRsvp6afh3HNh7Fg46qgSp0urjDZrFlbYx1u7NkxFjU9Qi/9Z/FhprU7r1ImWtBYd22YbqJl1M5lFREQSy7lktEhoG3U4ITHtDmyDe17k21NZwGRGI8Je9NsQVk19kM1VUiWj5bRqVWjxdPjhCfuODh8e5oiujOtkVq9e6Jm/qYuY3GH58tSS1+XLEz/LLCSkUZPX7bYLv0cUw4fDTTfBd99pAZeIiFSMHE5G6wMnAWcAxwAFpPB7RK2M1gT+AVwAG62aXwc8Clzmvr5nVdZQMroJbrgBBg6Eb78N2VYx2ZSMrV69ofoaJXn96afQVjWRevWSJ63bbQcffxy6YK1atfF9FZGMi4hI9ZVTyajZZkAXQjX0OMKi+DGEYfpXcF8R+VERk9HbgWuAW2IvWQz8IRZAf2CQO7em9ltUPiWjm2D+fNh5Z+jdG+68M9PRVKjCwtAsIGry+uOPEOVfoz/8Ab75Jnp1VUREJF7OJKNmI4DjCY32xxHWFI3CvVybhUdNRr8DHnTnngTnrgEud6dk+SzDlIxuoq5dYdIkWLAgTNisxn7/fUNi2qZN6Qu48vJg333DGrC2beHgg8PGVlp4JSIiZcmhZPRdQgL6Iu4/bfLjIiajq4AT3RmT4NwxwKvu1N3UYCqaktFNNHYsHHNMWNB09tmZjiZrlLaAa/vt4YILYMoUmDoVfv01HN9mm5CUHnxwSFAPOkjbqIqISEk5k4xWsKjJ6EzgQ3d6JTg3FGjtzn6VEN8mUTK6iQoLYc89Qzb1wQeZjiZrRFnAtW5d6Iw1ZQpMnhw+n322oaK6++4bV0/32SdUVUVEpPrKqWTUrAFwEWEVfdHC9onAo6kO10dNRrsTyrHjgBcJc0Z3AE4DOgBnuPNCKi9OByWjFeCBB+DKK8Pen61bZzqarFGeBVzLl4f/GYuS0ylTYMmScK5evTD8X5Sctm0LjRpV/u8hIiLZI2eSUbNdgPeA7YH/smEt0aGEHTo74P515MdFbe0UG47vBxwA1ALWAh8BfdwZm8KvkDZKRivAsmWw005w+ukwdGimo6lS3ENf1qLEdPLksFJ/zZpwvkmTDYlp27ZwwAGw2WYZDVlERCpRDiWjrwItgE64fx93fCfgTeBb3E+K/LhU+oyG91AD2A74KRvbOcVTMlpBLroInnoKFi4M+3hKpVm9GmbM2Lh6+u234VxeHuy338bD+7vuqsVRIiJVRQ4lo8uBHriXbEZudgrwBO5bRn5cqsloLlEyWkFmzgxZ0KBBcM01mY6m2lm8OCSlRdXTqVNhRax727bbllwc1aBBZuMVEZHyyaFk9P+A83F/McG504AhuG8V+XFKRiWSI46A77+HL78Mm9NLxhQtjiqqnk6eDLNnb1gctcceG1dPW7bU4igRkVyQQ8nov4HdCcP08+OONwPeAubgfnLkxykZlUhGjIAzzoDXX4fOnTMdjRSzfDlMm7bx/NMffwzn6tWDAw/ceP5pw4aZjVdERErKoWS0OWFRe2NgOhsWtrcGFgAdcZ8X+XFKRiWSNWugWbOwiub11zMdjZTBPcw1jW8t9fHHsHZtON+kycbVUy2OEhHJvJxJRgHMagO9gAOBhsAiYArwJO5rUnqUklGJrE8fuO22MFS/yy6ZjkZStGrVhsVRRUnqvHnhXF4etGq1cWupXXZJvDiqPG2tRESkbDmVjFagUpNRMxoBS9wpSPoAoz6wrzuTKiG+TaJktIJ9/32ojl55JdxTYmdYyUFFi6OKqqfTppVcHFVUQT3oIBg9uuyG/yIiUj5KRoufMNYBh7gzNfa9BvAV0MWdz+KuOxiY5E7NNMSbEiWjlaB7d3j77dDmqV69TEcjFWzdurAYKr56Gr84Ki8PChL89bRZsw1VVhERKZ/qmowmWxZdfIDOgOZAnUqLRrJffj788gs8/3ymI5FKULNm2Jr0ggtgyBCYNSv84x47Fm6/PXEiCmHIXkREpDySVUYLgbZxldGahF2X2rgzPe46VUarE/eQrdSpE/a2VMf1aqV5c5g/v+TxRo3CLA4RESk/VUZFojAL1dHp08M4rlQrd9yReHbGTz/BI49sGM4XERGJqqxkNNF/Wsr9nxsza2Jm481sjpl9ZmZXxI5vY2ZjzezL2J9bJ7i3g5nNiPusMrOu5Y1FNsHZZ8MWW8DgwZmORNLsrLPCYqVmzcLfS5o1gwcegHbt4JJL4NhjYcGCTEcpIiIVzuxGzBoWO3YEZpsXO9YCs0dTenQZw/TjgJ+LDgGnAO8Av8Rdug3QIcowvYVfoqG7TzezLYCPgK5AT+Bndx9gZtcDW7t77yTP2YawmKqxu68s7ToN01eSgQPhgw/gjTdC5rHDDjB+fFiKfd11mY5OMsA9JKlXXx3mnd5/P5x3nmZxiIikIquH6c3WAYfgPjX2vSawBjgQ9+lx1x0MTMI98vTNZJXRCUBNYPvYZzvgPSAv7tj2sWsmRHmZuy/yWMDu/iswB9gJOAkYFrtsGCFBTeZU4M1kiahUogMPhPfeC43whwwJiWj37uG4VEtmcNFFMHMm7L8/nH8+dOkCixZlOjIRkarFzIaa2RIzmxV3bJCZzTWzmWY2yswaJLgv4eh07FyZI9SUXNhe2rHUf6dMNb23sJXUBKAl8J27N4g794u7J/ofouj8OOA+dx+d4NyFwIUAtWvXbr169eoKjlyAkIAee2zo9VOzJrz4Yvgu1V5hIfzjH3DDDVC3Ljz0EJx5pqqkIiJliVIZNbMjgBXAU+7eMnbsGGCcuxeY2d0AxUeYSxuddvfZZjaQskaozQqBtsUqo2uBNpVZGa00ZlYfeAm40t2Xp3hvQ2Af4D+Jzrv7o+7ext3b5OXlbXqwkliHDtCrF/z+e+iSfv75cN99GzqmS7VVowZccUXY7WmPPcI801NPhSVLMh2ZiEjuc/cJbJhCWXRsjLsXNd+bTNgzvvh9pY1OQ+oj1BWq1GTUjEZmHJLgeCszXjLjMzPGmdEtlReaWS1CIjrc3V+OHV4cSzKLks1k/9nqDoxy97WpvFcq2Pjx8NJLcPPNsNVWsP32YcJgs2bQty8sXZrpCCXD/vhHmDgR7r477Ny0997hXxkRESlVnpl9GPe5sBzP6AW8meyC2Oj0/oS95AH+4O6LICStwA6l3FoXs3qY1QM2L3EsHN8s1YCTVUbvAO7dOHh2AyYCRxEy6q2BF83oGOVlZmbA48Acd78v7tSrQI/Yzz2AV5I85kzguSjvk0pSNEd05MiwV/2oUWFHpsGD4U9/gn79wqblV10Vjku1VbNmWNM2fXr4e8qpp8Kf/ww//1z2vSIi1VBB0ehu7JPaqnSzm4ACYHiSa8o9Og2MB36NfYoWs0+MO/YrYfF7SpKtpv8CeNCdh+KOPQycT2h8PzN27N/A5u4cXebLzA6PBf0pUBg7fCMhMx8JNAW+A05z95/NrA1wsbv/JXZ/c+C/QBN3L6QMWk1fSQYODIuVOnTYcCx+Nf3s2aEcNnx4GLM955xwfPfdMxezZNzatTBgAPTvD9ttB489BieckOmoRESyR9TV9LF8aHTRnNHYsR7AxUDH0hZ4x0anRwP/iS8KmtnnQHt3XxQboX7X3XcvdnMPUuE+rOyLYo9OkoyuAE5y5524Y98DX7nTLu7YicAj7jRKKcg0UDKaYfPnwz33hBX3q1fDKafA9ddD69aZjkwyaMYM6NEjrLzv2RP+/vcw00NEpLorbzJqZp2A+4B27v5jKfcYYT7oz+5+ZbFzg4ClcQuYtnH3tPVqTDZMv5K4cX8zWgANKdnG6RegRAsBEZo1C8uq588PS6vHjoU2beCYY0IlVdv1VEutWoUi+k03wdNPQ8uWMGZMpqMSEckNZvYc8AGwu5ktNLPzgYeALYCxsY2BHold28jM3ojdehhwDnBk3AZCnWPnBgBHm9mXwNGx7+UJbhfMbiKu7VSk25JURscAv7hzeuz7zUA/4FD39RNeMaMncIs7u5Qr8EqkymiWWb487Bl5332weDEcfHBIUrt0CcP5Uu1MmxaqpHPmhD6lgwaFzb1ERKqjrG56X5owrH86YU1PG2AdMBb34yM/Ikkyejhhoupswur2I4Hx7hxV7LrXgF/d+XN5fofKpGQ0S61aBU8+Geaefvst7LUX9O4dmlHWqpXp6CTNVq2CW26Be+8NxfQnnoD27TMdlYhI+uVMMhqa4p9KSECPIIy0OzAIuJdSpgqUptRylDvvAx0Ii41WE1bXb9TGyYztCQuRnkzlpVLN1a0LF18MX3wBzz4bGuf36AG77RY6pK/UxlrVSd26oSI6cWL4V6FDB7jySv1rICKSVcw2x+zPmL0G/AD8i9ACqg9wIGE3pjdSTUQhgzswpYMqoznCPexzf9dd8N//hp6lV14Jl14KDTQduTr57bewxu2hh8LfTYYNg0NKdDsWEamasroyaraCsJboW2AE8Bzus2LntiKsIWpPaMqfEk3Uk8wzg+OPh/ffhwkTQtuom24KvUp794Yffsh0hJImm28e1ryNGwdr1sDhh4d/BVatynRkIiLVnrFhL3qPfSrmwUnmjKbUtNSdIyskogqkymgO++ST0JRy5Mgwj/S88+Daa2HnnTMdmaTJr7+GTb0eeyxMK37qKXUFE5GqLcsro/UI24aeCRwL5BE2QHqOsEX7VMpZGU2WjBYSOum/C6wp60HunJbqyyubktEq4KuvwoTCJ5+EggI444xQKtt330xHJmny1lvwl7+EAvmNN4YdaGvXznRUIiIVL6uT0XhmDYDTgDOAdmyomN5HWMCU0pBmsmT0caArYSj/34TM9213ytz5KFsoGa1CFi2C+++Hhx+GFSvCsP4NN8Bhh2U6MkmDZcvgiitCdbRVqzCXVH8fEZGqJmeS0XhmOxKS0jOAgwitnd7G/bjIj0i2gMmMWoRS7BnAicDvhP1Mn3NnYvkjTw8lo1XQL7/A4MHwwAPw00/wpz+FpLRTpzD3VKq0V1+FCy8Me9v36ROK5Hl5mY5KRKRi5GQyGs+sBfBn4Azc94l8W9TV9GbUJSSkpwPHAT8Dd7jzcOrRpoeS0Srst9/g8cfDdqMLFsB++4Vl2KedBjVrZjo6qURLl0J+PowYEda6DRsGe+6Z6ahERDZdziej5ZRyaycz9gYuAi4FXnXn5MoIrCIoGa0G1q4NvUoHDIC5c2GXXeC660Lf0jp1Mh2dVKIXXoBLLgmzNm6/Ha66Sn8PEZHcltXJqNm5KVztuD8d+dFRklEzdmHDfIA9gPcJPaZedOenFIJLKyWj1UhhIbzySuhVOm0aNGwIf/tb2GNS+0tWWYsXh/0T/v3vMH34iSdCf1IRkVyU5cloIRvaOZU1L85xj1weSLaAaSfCkPwZQGtgGvA8MNKd/0V9QSYpGa2G3EOTygED4O23Q9P8v/4VLr88NNOXKscdhg+Hyy6D1avDLrOXXgo11EVZRHJMliejvxKS0VcI+eAESLKo3T1yApYsGS0AfgNei730m2QPcmd21Jemi5LRam7atJCUjhoV9py84ILQuLJp00xHJpXg++/DP+I33wxbig4dCs2bZzoqEZHosjwZrQucQChSdiasHRoBPI/7tE16dBl9RoskG8s3wN3JutlaSkYFCHNJ774bnnkmfD/77DCvVKteqhz3kIRedVX4+d57Q4KqRgsikguyOhmNZ1YfOJkwgn4UsJBQuHwa97kpPy5JMtoulQe5816qL69sSkZlIwsWhOzkscfg99+ha9fQFurAAzMdmVSw+fPh/PPhnXfg2GNhyBBo3DjTUYmIJJczyWg8s62BG4ErgddwT3lhe8qr6XOJklFJ6Kef4MEHwyboy5ZBx44hKT3ySJXQqpDCQnjkkbCLbK1a4R/5OefoH7GIZK+cSkbNGrPx2qKphN2XXkj1UZs8xd+MXc2yt9eoSAnbbQf9+8N334WtRmfPhqOOgoMPhpdfDlmM5LwaNcJCpk8+gX32Cd2+unYN24qKiEg5mG2P2aWYTQTmA+cQNkPaGfe25UlEoYxk1Iw8M9qacZoZbYqdO9CMF4G5QPfyvFwko7bYAq65Br79Fh59NOzudMopsPfe8OSToYep5Lxdd4V334X77oMxY8I/3hEjMh2ViEiOMfsP8D1wGfAOsBfurXAfgPu8TXp0kjmjzYDXgT2JLVIC3iRs8/QvQgK6FLgfeMidXzclkMqgYXpJybp18OKLoVfpJ59AkyYhWT3/fNg8N0ZNJLm5c6FnT5gyJWzW9c9/hkK5iEg2yOph+tBndAXwOckXtgfuB0V+dJJk9FmgPXAV8AnQDLgL2CL2c3/gXnd+j/qydFMyKuXiDv/5T0hKJ0wI2crll4d+pVtvnenoMmvgwLDgq0OHDcfGjw9ttK67LnNxpaCgIOwie+ut4R/nv/4Vhu9FRDIty5PRJ4mShBZxPy/yo5MkowuAW9x5Mu7YvsAM4Fp37o0cUIYoGZVNNmlSSEpHj4b69cN2P3XqhEVPOZyQldv48dC9O4wcGX7/4t9zyKefhnmkH38cun09+KD+riEimZXVyWglSpaMrgMOc2dy3LE8YA1wiDtT0hNi+SkZlQrz6aehV+nzz4fl2Hl5oUXU2WdvSMieew7atw8LoNwT/1kZ59L97M8+g8GDw+/8yis5mYgWWbsW7rgjfHbYIbSAOu64TEclItWVktHiJ0LT+4PdmRZ3rCawFmjtzsfpCbH8lIxKhfvmmzDGO2SIFjhBWB304ouw336ZjmSTTJ8O554b8uzzzw+LnbbcMtNRiUh1o2S0+ImQjM4hbAkarw0wG1gZf9CdyBNV00XJqFSaH34IVdF33oE//SkM25uFfkLF/0x0rDzXZMvzpkwJc2j33z9Uhd3hjDNCu6zddsv0P5lyW70a+vYN02IbNw47OXXsmOmoRKQ6UTJa/ITxRCoPcifyRNV0UTIqlaZoaP6SS+Dhh3N6qDolxeeIvvpqSEQLC8PKoPPPh1tuyentjiZPDivuP/8c8vPD7Aw1UxCRdFAyWgUpGZVKUYUW8aSstNX048aF3az+9a9QPf3rX+H663O2b9Lvv8NNN8Hf/w477xzazh5+eKajEpGqrromo5u8A5NItTNt2saJZ4cO4fu0acnvqwquu65kwt2hA9x2W9he9fPP4fTTw6TLnXcOQ/e/Zl0L4jJttln4Fd57L8xCOOIIuPpqeOIJaN485NvNm8Pw4ZmOVESqGzMbamZLzGxW3LFBZjbXzGaa2SgzaxD13tjxvmb2vZnNiH06V/bvsdH7VRkVkQr32WdhuH7UqFAdvfHGMKWhbt1MR5ayFSugd+/QIN8sJKdF6tULm3eddVbm4hORqiNKZdTMjiA0n3/K3VvGjh0DjHP3AjO7G8Dde0e5N3a8L7DC3e9JIdhTgZOBxkDJ/3NPoem9KqMiUvH23htefjksdmrVCv72t7C4aciQMLc0h9SvHzpZ7bDDxokowMqVYThfRCRd3H0C8HOxY2Pcvej/XCcTEsRI95ZLSF5HEnbpXAB8luAT/XGqjIpIpRs3Dm64AaZOhT/+MQzrn3pqGO/OETVqlExGIVRLCwvTH4+IVD1mtgb4NO7Qo+7+aILrmgOj46ubcedeA0a4+zOlvKPEvbHKaE9gOfAhcLW7/5Ik0AXA07jfWMavFEnu/JdARHLXkUeGZer//jfUqhXmlbZpA2++mTjDy0JNmyY+XqsWfPBBemMRkSqrwN3bxH1KJKLJmNlNQAGQ6oz2h4FdgFbAIihzl80tgHdSfEeplIyKSHqYwUknwSefwFNPhdX3nTtDu3bw/vuZjq5Md9wR5ojGq107HDv0UDjzTJg/PzOxiYiYWQ/gBOAsT3HY290Xu/s6dy8EHoMye8c/D3QqX6QlRU5GzTjVjGfNmGDG1OKfigpIRKq4mjXhnHNg7twwGfPLL8PGAccfDzNmZDq6Up11Vlis1KxZyKubNQuN8RcuhFtvDTuj7rFHmEOagw0ERCSHmVknoDdworuvLOv6BPc3jPvaDZhV2rUx7wCnYPYEZn/GrHOJTyrvj5I8m9EXuBX4hLD70pri16jpvYiUy2+/wUMPwYABoVqao7s5LVgQpsUOHw477hgqqT16hNxbRCSKiKvpnwPaA9sBi4E+wA1AHWBp7LLJ7n6xmTUChrh759LudffHzexpwhC9A/OAi9x9UZIgypop77hH/n+/qMnoAuBpdypkomq6KBkVySHLlsGgQaHT/OrV0KtXKDnm2G5OU6fCVVfBpEmhkcD990P79pmOSkRyQc40vTdrVuY17pEnLkUdpq/QiaoiIiU0aBDKiV9/DZdeGrY92nXX0G3+p58yHV1kBx0UpsA+/zz88kvYE6BbN/jqq0xHJpti+HBteCCynvv8Mj8piFoZfQT41Z1ryxt3JqgyKpLD5s2Dvn3h6afD5vBXXx1KjltumenIIvv991DovfPOUOy97LKwF0CDhHujSLYaPhwuvDD0lS2iDQ+kMuRMZbSIWR7QlMRN72dHfkzEZPQ04G7gPWAssKzkO3mj7OdYE+ApYEegkNA/6wEz2wYYATQnzFXonqi/lZk1BYYATQjzGjq7+7zS3qdkVKQKmD07ZHAvvwzbbht2c7r00pzazemHH8Kv8PjjsM020K8fXHQR5OVlOjKJonnzxJ0SmjULf2cSqSg5k4ya1QIeBHoQ5qqWVAlzRsucqOpOmS+NrdZq6O7TzWwL4COgK6HR6s/uPsDMrge2LmUbq3eBO9x9rJnVBwqTrRpTMipShUybFpaqjx0b5pHeeiucd15OZXSffBKKu+PHw557wr33wnHHZToqKUuyDQ/WrtUiNak4OZSM3kbI3a4j9DTNB34Dzib0K70M///27js+qjrd4/jnoUpRAYMalSK2lcuuIEXXgqCoWK4FK/ayKggrust69brXXbnLXgR0dQV1UVFRLKi4ICqiS7CwiqCCUsQGAhKKNIFASOC5fzxnzCRMyExIzsnMPO/Xa16Zds78TkL55tcerbCTMom35dsAACAASURBVCbZOaMHV3Brk8xJVDVfVT8N7m8EFgAHAucCTwdvexoLqKWISFugjqq+HRy/qTLbFzjn0lTnzjBlilVzOvBAGzdt2xZefDFtSiAddRT861+2939RkW2z2rMnzEupcJ4LQ1ERvPQSdO1afl0GVfujePPN9gvG9u3httG5CF0M/BkrCQrwMapjUD0N+ADLdUlLKoyq8n1Ft1Q+FH4uR9UBmAHsF9tCIPi6b4JDDgfWi8h4EflMRIaJyE6/j4rIjSIyS0RmFadZDWznXBK6d7eSRxMmQP36thVUx47wxhtpUc0ptvf/vHlw//0wY4aF1JtvhtWro26dW7nSqtW2bg0XX2z7yPbuDQ0alH5fgwZwyy0WVp9+2oqMHXAA9O1rvy/5fz8uw7UAvkJ1O7AVaBr32ljgglROllIFJhHqiNBGhLZlb6mdRxoDrwC3qupPSR5WBzgRGAh0xnpjryn7JlUdFSujVSeNhu+ccykQgXPOsU3yn3kGNmywTfO7dk2Lak5g1Ztuu832/O/b1xbDHHaYDd0XFkbduuyiatVqr7gCWrSwGSC//CVMmmQ/n+eeg8ceK13w4LHH4MEHYdw4WLXKelG7dbPiYqec4sHUZbx8ILYUcxHQNe61Q1I9WbJzRiucqJrMnFE7l9QFJgFvqer9wXMLgW6qmh/MK52mqkeUOe5YYIiqdgseXwkcq6r9yvssnzPqXJbYts1WBw0aZKuFzjjDtonq0CHqliVtwQIYONA6eA85xLZcPe88Cz+uemzdarM8RoyAWbNso4Zrr7Ve6sMPr9w5CwrgzTctpE6aZI+bN4deveCii6z6rfeTuPKk0ZzRJ4A1qN6OyK3AcGzIvhC4BHge1euTPV2yPaN3Y/VOrwcE6A9ci+09uhj4z+TaLgI8ASyIBdHARCzoEnydkODwmUBTEWkePD4ZqwblnMt29epZN9S331olp48+gqOPtiH8r76KunVJOfJIeP11mDzZZh/06mUzEj77LOqWZZ4lS6xa1kEHwTXXWGB8+GEbkn/ggcoHUbAtny64wELu6tXw8ss2hP/ss9Cjh/WY9uljc4e9x9Slsbuw3ZFA9QHgD0Ar4CjgIeCWVE6WbM/oQmAo8BRQBHRW5ZPgtaeBrarcVPF55ATgfeAL+HmF/n9j80bHYXtVLQEuUtW1ItIJ6KOqvwmOPxW4DwvEnwA3qupOpUljvGfUuSy1fj0MH27JYutW6+66+24bg00DxcU2DHz33bBmjQWmwYMhN7fCQ105VG2R0YgRNt0YbO7ub39rw+vV3QNdUGC/aMR6TDdvth7T88+3HtNu3bzH1KVRz2gVSzaMFgCnq/J+cP8cVd4JXjsNeE6VnOptauo8jDqX5VautB3nH33U0sbNN1uXWPPmFR9bA6xfbyH0wQet8/eOO2zv/7KLaUIxdKjtaNC9e8lzeXm25dbtt0fQoORs2mTTikeMsC1rc3Lghhusd7Jly2jaFAumL70Er71mwTQnp2Qo34Np9kq7MGo7HXXEFjSNRnUFIocCK7Fdk5KS7DB9lU5Udc65UOy3nyW5r76yJdEPPght2lhlp5+SXTsZnSZNbO7oggVw+um2cf4RR8Dzz0ewcUDnzra8PC/PHufl2ePOnUNuSHK++goGDCjZeqlBA6swu3Sp/X4SVRAFG8rv1ct+jqtXWz2HHj2s0tOpp1oP+I032pa6PpTvaiSRxoiMw0a6Hwf+FzggePWvwJ9SOp+qVngDfQJ0aHD/VtBi0OdAnwQtAH0imfOEfWvYsKE659zP5s9XveACVVDdZx/V4cNVCwqiblXSpk1T7dDBmn/ssaoffhjihxcWqo4dq7r33qp9+6rm5KhOnRpiAypWXKz62muqp59u36O6dVUvv9y+Tzt2RN26ihUUqI4fr9q7t2rjxiV/TH/zG9UpU1SLiqJuoatuwGatAfmpwhuMUlim0F2hnsIOhaOD165RmJvK+ZIdpt8fyFFlbvD4NuBCoAFWHnSQKjVuPNyH6Z1zCc2aZdWcpkyxrrNYNae6daNuWYV27LDtg+680zYO6N3b1mztVk/fjh02pWHJEus6XLp05/srV5bujq1XD048Ebp0gWOOsdv+++/29VXG2rXw5JO2COm770q2VbrhBuscT0dbtsBbb9kc09des+kG++xTMse0e/e0+OPqUpQ2w/QiPwIDUB2L7fleBHRC9VNEugMTUd0z6dMlE0bTlYdR59wuTZtmqe6jj2yTz0GDbOi5VkpbMEdi0ya4915bpwU2l/SOO6Bx4zJvVLXJp7sKmj/8YCWH4jVqZAu+WrSwpNuihU1uHDXK9ieaMsWeW7SoZCy5RQsLpbGA2rGjnaeazJljc0HHjrXw1rUr9O9vW2JlUlCLBdOXXoKJE0uC6Xnn2R9XD6aZI43C6GbgAlQnJwij5wBjUG2y65PEnS6VMBpsbv/zRFVVVohwKLBSlaQnqobFw6hzrkKq1vV0110wd66VQxo82Gp1psEmn0sXFvDgwKXMnrSUdnst5apuS2ifs5Ray+JCZ9l/B+vUsX2N4oNm2ftNm5a+/tgc0XHjLP3EHj/zjG3QOWMGfPyxfV20yI6pVQvatSsdUNu23a1i7kVF8OqrFkLff9/mgl5xBfTrZz+6TLd1a0mPaSyYNmtW0mN68skeTNNZGoXRacByVC9LEEbHADmonpn06ZIcpm8MjMbKOxVj1ZA6q/KpCOOAJaoMTP1qqpeHUedc0rZvhxdesCH7776DE06wlS4nnhhdm4qLYfnyXfdqrlmz02Gr6+xP/UNbsNd/lBM099sv9UCYymr61atLgunHH9tt3Tp7rVEj6NSp9PD+gQdWGPxXrLDtrh591L4lBx9sAfS66yw3Z6NYMI31mG7caME01mPqwTT9pFEYPQF4B6tD/xLwMLZo6QhsGmdXVGcmfbokw+go4EzgSmA6Voe0UxBGrwEGqtIutSupfh5GnXMpKyoqqeaUnw89e1oorepqTqoW2nYVNPPzbT5nvCZNEgfMli3Rg1rw0r8P5A9/rM+SJRZKhg2DQw+t2qanTBW++aZ07+ns2VY5C2z5eHzvaadOsNdeP5fpHDHCAldRkf04+ve3r7vRwZpxtm61mROxHtP4YHrRRVai1INpzZc2YRRA5HhgCHAsUBtQ4CPgdlSnp3SqJMPoj8AAVcaK8HN3bBBGuwMTVUl6ompYPIw65yqtoMBS0JAh1qv3q1/B734HV19d8p5d7bP500+7DprLlu1chH6PPcoNmj/f37Pif2q3bIG//c0y9LZtcMst8Mc/Wo6tMQoLbdJnfED9+msAVIT1uUfyfmEX3lhzDPMaHUPna9vR57d1d6s6UraIBdNYj+lPP1nvcazH1INpzZVWYTRGpAHQFFiPakGlTpFkGN0MXKDK5ARh9BxgjCo16Z85wMOoc64KbNhgq4SGD7f/5c84Ax56CD780Lro+va1VUOxsBn7WnYf09q1bZn3roJmTk6VzlPNz7cQ+uSTtuDlnnts/8qauqH60jlrmTJ4JqsnzaDdlo85rvYMmm3/0V5s0MBKvMZ6T7t0gdat02Jeb5QKC0v3mMYH01iPab16UbfSxaRlGK0CyYbRacByVS5LEEbHYNs+JT1RNSweRp1zVWblSts9ffz4xK83b77roJmbG1kK/Owz69SdNs3WD913nw1z1wSqMHWqdUJPnGjP/Vym8yRFvl9cuvf000/tlwKw73n88H7nztk7gTQJsWD60ktWEjUWTM89t6TH1INptGp0GBW5O4V3K6r/m/SpkwyjFU5UVSXpiaph8TDqnKtyAwbA3/8O55xj91u0sJXpkdToTJ6qBZCBA+Hbby2M3nefhdMobNxYUqZzwYIUynQWFcEXX5QOqF9+WbIH6uGHl14c9atfQf36oVxTOikstApP48aVBNMmTUp6THv08GAahRoeRncAW4DNQEVDEorqvkmfOtmtnUQod6KqKilNVA2Lh1HnXJWKbWfUty888kjJNkdpZNs2C4CDBtm2QDfdZMP3OTnhfP7ChbY5/VNPWQDq2NF6QS+5xKbMVsqGDVbIIBZOZ8yw5fdgiapDh9LD+4ce6sP7cWLBNNZjumGDBdNYj2mPHvbaXXfZLJSWLW33s8svj7rlmaeGh9FvgJbAv4AXgFdRrZK6yilvei/CzxNVVanURNWweBh1zlWZ8vbZTMNACvDjj/DnP9tWSY0b245W/ftXT2/Y9u3wxhsWgqdMscUzF19sn3fMMdWQC1VtgVh87+msWbYoDWyZeZcupQNqWGm8hisshHfeKekx3bDBOv23bbOfY0zDhlb/wANp1arRYRRApBNwKXAxkANMBp4HJqG6pdKn9QpMzjmXhFT22Uwj8+db9abJk63DcNgw6xGrioC4di2MHm09oYsWRVyms7jYLjY+oM6bV7J1Vps2pYf327cvmXqRoT/7isSC6SWX7Fw3AaBVK1i8OPRmZbQaH0bjiXTFgukFQENgIvAPVN9L+VTlhVERUpqoqkrSE1XD4mHUOeeSM3myhdL586FbN9saqn37yp1rzhzbcGDsWFtrVGPLdG7aBJ98Unp4f9kye61OHSvp1KWLVZgaNcrGqk85Je17xVNVq1bJlNyyfvjBfslwVSOZMCoio4GzgVWq2i54bhjwn8A24FvgWlVdn8yxwfPNgBeB1sBi4GJVXZdko+sBg4HbsJr0vZI6Lv4UuwijKU1UVSXpiaph8TDqnHPJKy62zHX33daree218Je/2EYAFSkqso0GRoyADz4oKdPZv7+tIUoby5eXVI2aMcN6PzcG1a5F4JBDYNUqeOUVm0yZBVq3hu+/T/xa3br2cx44MLrFcJkkyTDaFdgEjIkLo6cBU1W1WETuBVDV/0rm2OD5ocBaVR0iIncATRMdX+Zkx2M9oxcCewKTgEdQfTfpC45R1YQ30G9At4G+CXo16F7lvbem3ho2bKjOOedSs26d6u9/r1q3rmqjRqp/+YtqQYHqs8+qtmqlKmJfn31WNT9f9Z57VHNzVUG1TRvV++5TXbs26quoIsXFqvPmqY4erdqxo10kqLZoYd+YFSuibmG1e/ZZ1YYNSy4d7PH996v266faoIE9d9ZZqnl5qjt2RN3i9AVs1iTyDdaDObec184HxqZyLLAQyA3u5wILEx4PRysMVfheYavCPxUuVWiYTLvLbdOuL1Y7gQ4HXQJaADoe9CLQBrvzoWHdPIw651zlff216vnn2/8UzZqp1qtXOpDUrq1aq5bd79lTddIky24ZaepU1Zwc1bvuUt1rL9Wjj7YLr1tX9bLLVKdPz+gUlugXkZjVq+0XkubN7VvSqZPqiy+qFhVF1dr0VUVh9DXgilSOBdaXebxup2NhocI2hckK1yhUWSdlKls7JZyoqkrKE1XD4sP0zjm3+6ZNg9NOs6H4svbc0xaqZ3SZzvJ2Uhg+3Dbhj+1T1b69FUa47DJolB5rUKrSli0wZoztX/v113DwwXDbbXDddVn57agUEdkGfBH31ChVHZXgfa2BSRo31B48fxfQCeil5QS8RMeKyHpVbRL3eJ2qNi1z4A5gKzZ9s+LwWB37jJa0hVITVVVJeaJqWDyMOudc1ShvEYtIyYL0jFXRavpNm2y11siRtiH/3nvbhNubb4bDDouu3RHZvh1ee812Zvj3v63K0803236yoe+ikGaSXU1fTqC8GugDnKK7qBFfzrELgW6qmi8iucA0VT2izIF/SuliVO9J9q2pbnq/00RVVVKfqBoSD6POOVc1ylvE4tv7xFG11VsPPwwvv2wrwk47Dfr1g7POgtq1o25h6P79b+tA/uc/bQ/bK6+0XRt+8YuoW1YzVTaMikhP4H7gJFVdncqxwXPDgDVasoCpmaqGtm9ZrV29KMLRIgwV4Xtsx/0WWI/ovqpcWpODqHPOuaozeLBtdB6vYUN73gVE4MQT4fnnYelSK3M1b55t3NqmDfzf/8HqXeaEjHPccbbLwpdfWmfxs8/CkUdaNd333y9/yyhXPhF5HvgQOEJElonI9cAIrKPwbRGZLSKPBu89QETeqOBYsAqbp4rI18CpwePQ7Gprp4XAwcBUrOzTeFWqpOxTWLxn1Dnnqs7YsV4SMmXFxTBxog3hT51q3YMXX2y9pdVSfqpmW7XKvhUjR8KaNfYtGDgQzj8/KzuOd5JWm95XoYr2GU16oqr6PqPOOedc+RYssCH8p5+2vUs7dLBQ2rv3zt3OGa6gwNZ93X8/fPutbd/6u9/BNddk3beiFA+jZV8QUpqoqkrSE1XD4mHUOedcjbNxo41Xjxxpw/hNm9oYdt++VpM1i2zfbvNJhw2zGgP77GP5vF8/2LfGdXFVPw+jGcjDqHPOuRpL1SZOjhxpEyuLi6FnT1t6fuaZWTVurQrTp1sonTgR9tgDrr7aFjtl04YEHkYzkIdR55xzaWH5cnjsMavHuny5bV/Qpw9cfz3k5ETdulB9+aUN348ZA9u2wXnn2bzS446LumXVz8NoBvIw6pxzLq0UFcGECdZbOm0a1K8Pl1xivaVdumTVgqeVK2HECPtWrFtnYXTgQFuJn6mdxh5GM5CHUeecc2lr3jxb8DRmjG2s37GjTaa89FJo0CDq1oVm82YYPdp6SxcvtmH73/8erroq874NHkYzkIdR55xzae+nn0oWPM2fD82aWY3NPn1sGXqWKC62qbXDhlkJ2ubNoX9/6zTOlJkMHkYzkIdR55xzGUMV3n3XQumrr1od1p49rbe0Z8/MHbsuQxXee89C6euvW+/otdfa1lDpns09jGYgD6POOecy0g8/lCx4ys+Hgw+2raGuu872R8oS8+fDffdZx3FREfTqBX/4g22mn448jGYgD6POOecyWlGR9ZKOHGndhfXr25zSfv2gc+eoWxea/Hx46CF45BFYv96qsg4cCGefDbV2Wfi8ZsnWMJpGPyLnnHPOlVK3rpUXffdd+Pxz6xl9+WVbed+li1V72ro16lZWu9xc+OtfrVTtAw/Y13PPhbZtrQO50t+CoUMhL6/0c3l59ryrMh5GnXPOuUzwy1/a6vvly62bcONGq6950EFw++2waFHULax2e+4JAwbAN9/Ac89ZadEbb7RtWwcPhrVrUzxh584W9mOBNC/PHmdRr3MYfJjeOeecy0SqFp4efthqbu7YYZWd+vWD009Pr/HrSop9C4YPhzfftHB6/fVw2202zXaXduyANWts39eBA+179/bbMG4cdO9eLe3N1mF6D6POOedcplu2zBY7jRplu8m3aVOy4KlZs6hbF4q5cy2Uvjq2gH2359O7+wquOn0FhzZaYZNOV6ywW+z+ypWwfXvpk/zP/8CgQdXWRg+jGcjDqHPOORdn2zbbrHPkSPjgAysC37u39ZZ27Bh16ypv+3b48cfSoXJFOSFz48adDtdatWDffZHcXNh//5Jbbq6d98EH4cor4YUXvGe0GoQaRkWkBTAG2B/YAYxS1QdFpBnwItAaWAxcrKrrEhy/HfgieLhEVc/Z1ed5GHXOOefK8fnnNoT/zDNQUGALnvr1szmRe+wRdevMpk3lh8r4x6tW2bB6WXvuWRIq4wPm/vuzea/9GT99f/72Qi5zlufwi7a1GTgQLrvMNiUASuaIxgJo2cdVzMNoGB8mkgvkquqnIrIn8AlwHnANsFZVh4jIHUBTVf2vBMdvUtXGyX6eh1HnnHOuAhs22Kr7hx+GhQutnNH111uFp9atbeV4586lw1deHsycaQujUlVcDKtXJxcyE/0fXrs27Ldf6YCZKHDutx80qjjXFRVZthw2DObMsUNvuQVuugmaPjaUdzZ05jdju7NkCbRsCY9fnkePvSt57RXwMBrFh4tMAEYEt26qmh8E1mmqekSC93sYdc4556qDKkydakP4EybY47PPhuOPt8mWu+odVLXh7/JCZfz91avt/WXtvXe5vZilHu+zT7UsvlKFd96xS50yBRo3hhNOsF2ztmwpeV/Dhjb19vLLq7wJHkZD/2CR1sB7QDtsyL1J3GvrVLVpgmOKgdlAMTBEVf+5q8/wMOqcc85VwtKl8I9/2Cadq1bBAQdYD2r37hZYTz3VeijjQ2Z8YoupU6f8UFm2F7NBg/Cvsxxz5lhlp2eeSfx6q1aweHHVf66H0TA/VKQx8C4wWFXHi8j6JMPoAaq6XETaAFOBU1T12zLvuRG4EaBevXodCwsLq/VanHPOuYxVWAivvGJD+NOnlzzftGlyvZhNm6b1FlK1aiXuxBVJPEV1d2VrGK0T9geKSF3gFWCsqo4Pnl4pIrlxw/SrEh2rqsuDr9+JyDSgA/BtmfeMAkaB9YxWz1U455xzWaB+fVvRk5sLF14IV1xhu8lX44rymqRlS/j++8TPu6oT6q8rIiLAE8ACVb0/7qWJwNXB/auBCQmObSoi9YP7OcDxwPzqbbFzzjmX5WJzRF9+2bY4GjeudFWiDDZ4sM0RjdewoT3vqk7YfefHA1cCJ4vI7OB2JjAEOFVEvgZODR4jIp1E5PHg2COBWSIyB8jD5ox6GHXOOeeq08yZpXtCu3e3xzNnRtuuEFx+uS1WatXKhuZbtaq+xUvZzDe9d84555yrAbJ1zmj6zip2zjnnnHNpz8Ooc84555yLjIdR55xzzrk0ISKjRWSViMyNe26YiHwpIp+LyKsi0qScY3uKyEIR+SaoeBl7/ikRWRS3nqd9GNcS42HUOeeccy59PAX0LPPc20A7Vf0V8BVwZ9mDRKQ2MBI4A2gL9BaRtnFv+YOqtg9us6ul5eXwMOqcc845lyZU9T1gbZnnpqhqcfDwI+CgBId2Ab5R1e9UdRvwAnButTY2SR5GnXPOOedqhjoiMivudmMlznEd8GaC5w8ElsY9XhY8FzM4GOb/W2xf97CEXoHJOeecc84lVKyqnSp7sIjcBRQDYxO9nOC52P6edwIrgHpYFcv/AgZVth2p8p5R55xzzrk0JyJXA2cDl2viTeSXAS3iHh8ExMqs56spBJ7EhvRDk9E9owUFBSoiW0L4qDrYbyLZKJuvHbL7+v3as1c2X382Xztk9/WHce0NKnOQiPTEejNPUtWCct42EzhMRA4GfgAuBS4Ljs9V1fygbPt5wNxyzlEtMjqMqmooPb8iMmt3utXTWTZfO2T39fu1Z+e1Q3ZffzZfO2T39deUaxeR54FuQI6ILAP+hA2z1wfetjzJR6raR0QOAB5X1TNVtVhE+gNvAbWB0ao6LzjtWBFpjg3lzwb6hHlNGR1GnXPOOecyiar2TvD0E+W8dzlwZtzjN4A3Erzv5CprYCX4nFHnnHPOORcZD6NVY1TUDYhQNl87ZPf1+7Vnr2y+/my+dsju68/ma69WknjBlXPOOeecc9XPe0adc84551xkPIw655xzzrnIeBitJBFpISJ5IrJAROaJyICo2xQmEdlDRD4WkTnB9d8TdZvCJiK1ReQzEZkUdVvCJiKLReQLEZktIrOibk+YRKSJiLwsIl8Gf/9/HXWbwiAiRwQ/79jtJxG5Nep2hUlEbgv+vZsrIs+LyB5RtyksIjIguO552fBzF5HRIrJKRObGPddMRN4Wka+Dr02jbGMm8TBaecXA71X1SOBYoJ+ItI24TWEqBE5W1aOA9kBPETk24jaFbQCwIOpGRKi7qravCfvuhexBYLKq/gI4iiz5M6CqC4Ofd3ugI1AAvBpxs0IjIgcCtwCdVLUdtk/jpdG2Khwi0g64AavKcxRwtogcFm2rqt1TQM8yz90B/EtVDwP+FTx2VcDDaCUFpbM+De5vxP5DOjDaVoUnKBu2KXhYN7hlzWo4ETkIOAt4POq2uPCIyF5AV4I9/VR1m6quj7ZVkTgF+FZVv4+6ISGrAzQQkTpAQ4JSilngSGwT9QJVLQbeBc6PuE3VSlXfA9aWefpc4Ong/tNYpSJXBTyMVgERaQ10AGZE25JwBcPUs4FVwNuqmk3X/wBwO7Aj6oZERIEpIvKJiNwYdWNC1AZYDTwZTNF4XEQaRd2oCFwKPB91I8Kkqj8Aw4ElQD6wQVWnRNuq0MwFuorIPiLSENtEvUUFx2Si/VQ1H6xDCtg34vZkDA+ju0lEGgOvALeq6k9RtydMqro9GLI7COgSDOVkPBE5G1ilqp9E3ZYIHa+qRwNnYFNUukbdoJDUAY4GHlHVDsBmsmyoTkTqAecAL0XdljAF8wPPBQ4GDgAaicgV0bYqHKq6ALgXeBuYDMwhe+vTu2rgYXQ3iEhdLIiOVdXxUbcnKsEw5TR2nl+TqY4HzhGRxcALwMki8my0TQpXUGIOVV2FzRvsEm2LQrMMWBY3CvAyFk6zyRnAp6q6MuqGhKwHsEhVV6tqETAeOC7iNoVGVZ9Q1aNVtSs2fP111G2KwEoRyQUIvq6KuD0Zw8NoJYmIYPPGFqjq/VG3J2wi0lxEmgT3G2D/UH8ZbavCoap3qupBqtoaG66cqqpZ0UMCICKNRGTP2H3gNGwYL+Op6gpgqYgcETx1CjA/wiZFoTdZNkQfWAIcKyINg3//TyFLFq8BiMi+wdeWQC+y88/ARODq4P7VwIQI25JR6kTdgDR2PHAl8EUwbxLgv1X1jQjbFKZc4GkRqY39UjNOVbNui6MstR/wqv1/TB3gOVWdHG2TQvVbYGwwXP0dcG3E7QlNMF/wVOCmqNsSNlWdISIvA59iQ9SfkV3lIV8RkX2AIqCfqq6LukHVSUSeB7oBOSKyDPgTMAQYJyLXY7+cXBRdCzOLlwN1zjnnnHOR8WF655xzzjkXGQ+jzjnnnHMuMh5GnXPOOedcZDyMOuecc865yHgYdc4555xzkfEw6pyrdiKiInJf3OOBIvLnKjr3UyJyYVWcq4LPuUhEFohIXpnnW4vIZXGPO4nI36u7PRURkU3B1wOCLYkQkfYicma0LXPOudI8jDrnwlAI9BKRnKgbEi/YJzdZ1wM3q2r3Ms+3Bn4Oo6o6S1VvqYLmN2n2CQAAA+FJREFUVQlVXa6qsbDeHqsr7pxzNYaHUedcGIqxDcJvK/tC2Z7NuB69biLyroiME5GvRGSIiFwuIh+LyBcickjcaXqIyPvB+84Ojq8tIsNEZKaIfC4iN8WdN09EngO+SNCe3sH554rIvcFzdwMnAI+KyLAyhwwBThSR2SJyW3D+ScFxfxaRp0VkiogsFpFeIjI0OP/koKQwItIxuNZPROStWMnBMu26KGjTHBF5L3juGhGZEJxroYj8KcFxrYPj6gGDgEuCtl4iIicF92eLyGexylrOORcmr8DknAvLSOBzERmawjFHAUditbC/Ax5X1S4iMgCrhHRr8L7WwEnAIUCeiBwKXAVsUNXOIlIfmC4iU4L3dwHaqeqi+A8TkQOAe4GOwDpgioicp6qDRORkYKCqzirTxjuC52MhuFuZ1w8BugNtgQ+BC1T1dhF5FThLRF4HHgLOVdXVInIJMBi4rsx57gZOV9UfYqV4468FKABmisjrCdqIqm4LQnUnVe0ftPU1rJrOdBFpDGwte5xzzlU37xl1zoVCVX8CxgCpDGHPVNV8VS0EvgViYfILLIDGjFPVHar6NRZafwGcBlwVlOudAewDHBa8/+OyQTTQGZimqqtVtRgYC3RNob2JvKmqRUGbawOx0qmxazgCC5NvB239I3BQgvNMB54SkRuC88S8raprVHULMB7rwU3WdOB+EbkFaBJcs3POhcp7Rp1zYXoAq+39ZNxzxQS/GIsVvK8X91ph3P0dcY93UPrfr7J1jRUQ4Leq+lb8C0HP5eZy2icVXkHqCgFUdYeIFGlJDebYNQgwT1V/vauTqGofETkGOAuYLSLtYy+VfWuyDVPVIUHP7JnARyLSQ1W/TPZ455yrCt4z6pwLjaquBcZhi4FiFmPD4gDnAnUrceqLRKRWMI+0DbAQeAvoGzcv83ARaVTBeWYAJ4lITrC4qTfwbgXHbAR2Z67lQqC5iPw6aGddEfmPsm8SkUNUdYaq3g38CLQIXjpVRJqJSAPgPKy3M6m2Buf8QlXvBWZhPcrOORcqD6POubDdB8Svqn8MC4AfA8dQfq/lrizEQuObQB9V3Qo8DswHPhWRucA/qGA0SFXzgTuBPGAO8KmqTqjgsz8HioOFRTst0KqIqm4DLgTuFZE5wGzguARvHRZbWAW8F7QP4APgmeC4VxLNF42TB7SNLWACbo0tigK2YN8/55wLlZSMGDnnnEsnInINcQuSnHMuHXnPqHPOOeeci4z3jDrnnHPOuch4z6hzzjnnnIuMh1HnnHPOORcZD6POOeeccy4yHkadc84551xkPIw655xzzrnI/D/JmvHPleoczgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax1 = plt.subplots(figsize=(10,5))\n",
    "ax1.plot(splits, rmse_split, 'bo-')\n",
    "\n",
    "ax2 = ax1.twinx()\n",
    "ax2.plot(splits, mae_split, 'rx-')\n",
    "ax1.set_xlabel(\"Number of time splits\")\n",
    "ax1.set_ylabel(\"Mean RMSE over time splits\", color='b', fontsize=15)\n",
    "ax2.set_ylabel(\"Mean MAE over time splits\", color='r', fontsize=15);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The computing has been long but the outcome shows how cross-validation improves the model performance. However looking at the scale of the RMSE reduction, one can ask if such a small gain is really worth the trouble as it might be more relevant to wonder about the behaviour of the model itself."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Saving the objects:\n",
    "with open(path + '/data/objs_cv_lr.pkl', 'wb') as f :  \n",
    "    pickle.dump([splits, rmse_split, mae_split], f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 8) Other Ridge models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the exploratory phase we aknowledged the influence of the departure airport and the airline carrying the flight. We will now try to observe if selecting one airport or one airline or both improves the Ridge regression performances."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 8.1) Same data but ONE airport, ALL airlines"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1) Loading dataframe (pickle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "366245 entries for this airport\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TAIL_NUM</th>\n",
       "      <th>UNIQUE_CARRIER</th>\n",
       "      <th>ORIGIN_CITY_NAME</th>\n",
       "      <th>ARR_DELAY_NEW</th>\n",
       "      <th>DISTANCE</th>\n",
       "      <th>DATE_TIME</th>\n",
       "      <th>TIME_SECONDS</th>\n",
       "      <th>ARR_TIME_SECS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>126697</th>\n",
       "      <td>N506NK</td>\n",
       "      <td>NK</td>\n",
       "      <td>Atlanta, GA</td>\n",
       "      <td>8.0</td>\n",
       "      <td>404.0</td>\n",
       "      <td>2016-01-01 06:00:00</td>\n",
       "      <td>21600</td>\n",
       "      <td>27120.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>332324</th>\n",
       "      <td>N947FR</td>\n",
       "      <td>F9</td>\n",
       "      <td>Atlanta, GA</td>\n",
       "      <td>0.0</td>\n",
       "      <td>666.0</td>\n",
       "      <td>2016-01-01 06:05:00</td>\n",
       "      <td>21900</td>\n",
       "      <td>29760.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       TAIL_NUM UNIQUE_CARRIER ORIGIN_CITY_NAME  ARR_DELAY_NEW  DISTANCE  \\\n",
       "126697   N506NK             NK      Atlanta, GA            8.0     404.0   \n",
       "332324   N947FR             F9      Atlanta, GA            0.0     666.0   \n",
       "\n",
       "                 DATE_TIME  TIME_SECONDS  ARR_TIME_SECS  \n",
       "126697 2016-01-01 06:00:00         21600        27120.0  \n",
       "332324 2016-01-01 06:05:00         21900        29760.0  "
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open(path + '/data/data_2016_1_aiport.pkl', 'rb') as f :\n",
    "    data_1_airport = pickle.load(f)\n",
    "print(\"%i entries for this airport\" % data_1_airport.shape[0])\n",
    "data_1_airport.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2) Encoding airlines"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [],
   "source": [
    "le = preprocessing.LabelEncoder()\n",
    "label_airlines = le.fit_transform(data_1_airport.UNIQUE_CARRIER)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "3) One Hot encoding airlines from label encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "oh_airlines = pd.get_dummies(label_airlines)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "4) Standardise numerical features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "std_scale = preprocessing.StandardScaler().fit(data_1_airport.loc[:, ['TIME_SECONDS', 'ARR_TIME_SECS', 'DISTANCE']].values)\n",
    "data_num_1_airport = pd.DataFrame(std_scale.transform(data_1_airport.loc[:, ['TIME_SECONDS', 'ARR_TIME_SECS', 'DISTANCE']].values), columns=['DEP_TIME_SECS_STD', 'ARR_TIME_SECS_STD', 'DISTANCE_STD'])\n",
    "# Reindex\n",
    "data_num_1_airport.index = oh_airlines.index\n",
    "# Concat\n",
    "data_1_airport_reg = pd.concat((oh_airlines, data_num_1_airport), axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "5) Reindex dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_1_airport.index = oh_airlines.index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "6) Concatenate everything into processed data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>DEP_TIME_SECS_STD</th>\n",
       "      <th>ARR_TIME_SECS_STD</th>\n",
       "      <th>DISTANCE_STD</th>\n",
       "      <th>ARR_DELAY_NEW</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-1.887855</td>\n",
       "      <td>-1.635402</td>\n",
       "      <td>-0.524538</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-1.870178</td>\n",
       "      <td>-1.493882</td>\n",
       "      <td>0.034261</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0  1  2  3  4  5  6  7  8  DEP_TIME_SECS_STD  ARR_TIME_SECS_STD  \\\n",
       "0  0  0  0  0  0  1  0  0  0          -1.887855          -1.635402   \n",
       "1  0  0  0  0  1  0  0  0  0          -1.870178          -1.493882   \n",
       "\n",
       "   DISTANCE_STD  ARR_DELAY_NEW  \n",
       "0     -0.524538            8.0  \n",
       "1      0.034261            0.0  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data_1_airport_reg = pd.concat((data_1_airport_reg, data_1_airport.loc[:, 'ARR_DELAY_NEW']), axis=1)\n",
    "display(data_1_airport_reg.head(2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Target : arrival delay\n",
    "y = data_1_airport_reg.loc[:, 'ARR_DELAY_NEW']\n",
    "X = data_1_airport_reg.loc[:, data_1_airport_reg.columns  != 'ARR_DELAY_NEW']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "7) Cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cross validation with 2 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 22.76, MAE : 12.19, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 16.11, MAE : 10.09, R² : -0.02\n",
      "Mean RMSE 19.44\n",
      "Cross validation with 3 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 18.81, MAE : 10.39, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 22.32, MAE : 12.27, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 16.19, MAE : 9.91, R² : -0.02\n",
      "Mean RMSE 19.11\n",
      "Cross validation with 4 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 16.67, MAE : 9.37, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 25.28, MAE : 13.56, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 17.03, MAE : 10.50, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 16.81, MAE : 9.99, R² : -0.01\n",
      "Mean RMSE 18.95\n",
      "Cross validation with 5 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 16.66, MAE : 9.57, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 20.13, MAE : 10.88, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 25.02, MAE : 13.77, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 14.74, MAE : 9.61, R² : -0.06\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 17.33, MAE : 10.14, R² : -0.00\n",
      "Mean RMSE 18.78\n",
      "Cross validation with 6 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 16.42, MAE : 9.53, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 16.82, MAE : 9.45, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 25.63, MAE : 13.80, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 21.13, MAE : 12.09, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 14.20, MAE : 9.28, R² : -0.07\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 17.87, MAE : 10.32, R² : -0.00\n",
      "Mean RMSE 18.68\n",
      "Cross validation with 7 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 17.44, MAE : 9.96, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 15.84, MAE : 9.19, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 21.39, MAE : 11.48, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 26.86, MAE : 14.84, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 16.72, MAE : 10.46, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 13.53, MAE : 8.94, R² : -0.09\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 18.48, MAE : 10.56, R² : -0.00\n",
      "Mean RMSE 18.61\n",
      "Cross validation with 8 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 17.50, MAE : 9.75, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 16.84, MAE : 9.49, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 18.20, MAE : 9.95, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 26.02, MAE : 14.17, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 23.18, MAE : 13.16, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 15.27, MAE : 9.83, R² : -0.05\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 13.60, MAE : 8.83, R² : -0.08\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 8 : RMSE : 18.92, MAE : 10.76, R² : -0.00\n",
      "Mean RMSE 18.69\n",
      "Cross validation with 9 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 17.27, MAE : 9.64, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 16.33, MAE : 9.21, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 17.01, MAE : 9.45, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 22.58, MAE : 12.08, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 27.57, MAE : 15.30, R² : -0.05\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 19.07, MAE : 11.37, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 14.67, MAE : 9.50, R² : -0.06\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 8 : RMSE : 13.90, MAE : 8.84, R² : -0.06\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 9 : RMSE : 19.30, MAE : 10.94, R² : -0.00\n",
      "Mean RMSE 18.63\n",
      "Cross validation with 10 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 17.71, MAE : 9.92, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 16.79, MAE : 9.53, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 16.32, MAE : 9.29, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 20.29, MAE : 10.87, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 25.16, MAE : 13.79, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 26.40, MAE : 14.89, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 15.15, MAE : 9.92, R² : -0.06\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 8 : RMSE : 14.19, MAE : 9.26, R² : -0.06\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 9 : RMSE : 13.60, MAE : 8.69, R² : -0.07\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 10 : RMSE : 19.95, MAE : 11.25, R² : -0.01\n",
      "Mean RMSE 18.56\n"
     ]
    }
   ],
   "source": [
    "splits = [2, 3, 4, 5, 6, 7, 8, 9, 10]\n",
    "mae_split = []\n",
    "rmse_split = []\n",
    "r2_split = []\n",
    "\n",
    "# Model\n",
    "ridge = linear_model.Ridge()\n",
    "\n",
    "for split in splits :\n",
    "    print(\"Cross validation with %i time splits\" % split)\n",
    "    errors_splits, alphas_splits = time_cross_validation(ridge, X, y, 5, split)\n",
    "    \n",
    "    errors_array = np.array(errors_splits)\n",
    "    \n",
    "    mae_split.append(errors_array[:,0].mean())\n",
    "    rmse_split.append(errors_array[:,1].mean())\n",
    "    r2_split.append(errors_array[:,2].mean())\n",
    "    \n",
    "    print(\"Mean RMSE %.2f\" % errors_array[:,1].mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax1 = plt.subplots(figsize=(10,5))\n",
    "ax1.plot(splits, rmse_split, 'bo-')\n",
    "\n",
    "ax2 = ax1.twinx()\n",
    "ax2.plot(splits, mae_split, 'rx-')\n",
    "ax1.set_xlabel(\"Number of time splits\", fontsize=15)\n",
    "ax1.set_ylabel(\"Mean RMSE over time splits\", color='b', fontsize=15)\n",
    "ax2.set_ylabel(\"Mean MAE over time splits\", color='r', fontsize=15);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The CV displays a best result for 7 time splits, but the model seems worse than the main Ridge Regression with all airports and all airlines."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 8.2) Same data but ONE airline, ONE airport"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1) Loading dataframe (pickle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "80917 entries for this airport\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TAIL_NUM</th>\n",
       "      <th>ORIGIN_CITY_NAME</th>\n",
       "      <th>ARR_DELAY_NEW</th>\n",
       "      <th>DISTANCE</th>\n",
       "      <th>TIME_SECONDS</th>\n",
       "      <th>ARR_TIME_SECS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>339550</th>\n",
       "      <td>N707SA</td>\n",
       "      <td>Chicago, IL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1521.0</td>\n",
       "      <td>22500</td>\n",
       "      <td>30240.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>339642</th>\n",
       "      <td>N8645A</td>\n",
       "      <td>Chicago, IL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>997.0</td>\n",
       "      <td>25200</td>\n",
       "      <td>39360.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       TAIL_NUM ORIGIN_CITY_NAME  ARR_DELAY_NEW  DISTANCE  TIME_SECONDS  \\\n",
       "339550   N707SA      Chicago, IL            0.0    1521.0         22500   \n",
       "339642   N8645A      Chicago, IL            0.0     997.0         25200   \n",
       "\n",
       "        ARR_TIME_SECS  \n",
       "339550        30240.0  \n",
       "339642        39360.0  "
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open(path + '/data/data_2016_1_airline_1_aiport.pkl', 'rb') as f :\n",
    "    data_1_airline_1_aiport = pickle.load(f)\n",
    "print(\"%i entries for this airport\" % data_1_airline_1_aiport.shape[0])\n",
    "data_1_airline_1_aiport.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2) Standardise numerical features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [],
   "source": [
    "std_scale = preprocessing.StandardScaler().fit(data_1_airline_1_aiport.loc[:, ['TIME_SECONDS', 'ARR_TIME_SECS', 'DISTANCE']].values)\n",
    "data_num_1_airline_1_airport = pd.DataFrame(std_scale.transform(data_1_airline_1_aiport.loc[:, ['TIME_SECONDS', 'ARR_TIME_SECS', 'DISTANCE']].values), columns=['DEP_TIME_SECS_STD', 'ARR_TIME_SECS_STD', 'DISTANCE_STD'])\n",
    "# Reindex\n",
    "data_num_1_airline_1_airport.index = data_1_airline_1_aiport.index\n",
    "# Concat\n",
    "data_1_airport_reg = pd.concat((data_1_airline_1_aiport['ARR_DELAY_NEW'], data_num_1_airline_1_airport), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ARR_DELAY_NEW</th>\n",
       "      <th>DEP_TIME_SECS_STD</th>\n",
       "      <th>ARR_TIME_SECS_STD</th>\n",
       "      <th>DISTANCE_STD</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>339550</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.575404</td>\n",
       "      <td>-1.140637</td>\n",
       "      <td>1.489495</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>339642</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.426590</td>\n",
       "      <td>-0.718927</td>\n",
       "      <td>0.377094</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        ARR_DELAY_NEW  DEP_TIME_SECS_STD  ARR_TIME_SECS_STD  DISTANCE_STD\n",
       "339550            0.0          -1.575404          -1.140637      1.489495\n",
       "339642            0.0          -1.426590          -0.718927      0.377094"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_1_airport_reg.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Target : arrival delay\n",
    "y = data_1_airline_1_aiport.loc[:, 'ARR_DELAY_NEW']\n",
    "X = data_1_airport_reg.loc[:, data_1_airport_reg.columns  != 'ARR_DELAY_NEW']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "3) Cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cross validation with 2 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 22.35, MAE : 12.24, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 17.28, MAE : 10.98, R² : -0.00\n",
      "Mean RMSE 19.82\n",
      "Cross validation with 3 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 18.40, MAE : 10.31, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 21.72, MAE : 12.31, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 17.72, MAE : 11.07, R² : -0.00\n",
      "Mean RMSE 19.28\n",
      "Cross validation with 4 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 18.20, MAE : 9.82, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 23.54, MAE : 13.16, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 16.87, MAE : 10.67, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 18.55, MAE : 11.37, R² : -0.00\n",
      "Mean RMSE 19.29\n",
      "Cross validation with 5 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 16.98, MAE : 9.18, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 20.15, MAE : 11.04, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 24.16, MAE : 13.73, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 15.05, MAE : 10.04, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 19.27, MAE : 11.67, R² : -0.00\n",
      "Mean RMSE 19.12\n",
      "Cross validation with 6 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 16.31, MAE : 8.98, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 18.92, MAE : 10.34, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 25.13, MAE : 14.12, R² : -0.06\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 17.92, MAE : 11.11, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 15.41, MAE : 10.09, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 19.84, MAE : 11.94, R² : -0.00\n",
      "Mean RMSE 18.92\n",
      "Cross validation with 7 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 16.82, MAE : 9.43, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 15.37, MAE : 9.04, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 21.01, MAE : 11.57, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 25.74, MAE : 14.68, R² : -0.06\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 16.84, MAE : 10.65, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 14.44, MAE : 9.75, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 20.50, MAE : 12.25, R² : -0.01\n",
      "Mean RMSE 18.67\n",
      "Cross validation with 8 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 15.80, MAE : 8.85, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 16.16, MAE : 9.25, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 20.28, MAE : 11.07, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 25.35, MAE : 14.20, R² : -0.06\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 20.67, MAE : 12.41, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 15.23, MAE : 10.01, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 15.55, MAE : 10.09, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 8 : RMSE : 20.58, MAE : 12.33, R² : -0.01\n",
      "Mean RMSE 18.70\n",
      "Cross validation with 9 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 15.78, MAE : 8.52, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 17.27, MAE : 9.45, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 19.01, MAE : 10.45, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 19.89, MAE : 11.24, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 26.58, MAE : 15.20, R² : -0.07\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 18.47, MAE : 11.29, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 15.10, MAE : 10.02, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 8 : RMSE : 15.62, MAE : 10.04, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 9 : RMSE : 21.10, MAE : 12.62, R² : -0.02\n",
      "Mean RMSE 18.76\n",
      "Cross validation with 10 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 15.16, MAE : 8.19, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 17.93, MAE : 9.75, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 16.54, MAE : 9.42, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 20.45, MAE : 11.27, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 23.04, MAE : 12.92, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 25.93, MAE : 15.09, R² : -0.06\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 15.24, MAE : 10.11, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 8 : RMSE : 14.71, MAE : 9.83, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 9 : RMSE : 15.53, MAE : 9.95, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 10 : RMSE : 21.94, MAE : 13.07, R² : -0.03\n",
      "Mean RMSE 18.65\n"
     ]
    }
   ],
   "source": [
    "splits = [2, 3, 4, 5, 6, 7, 8, 9, 10]\n",
    "mae_split = []\n",
    "rmse_split = []\n",
    "r2_split = []\n",
    "\n",
    "# Model\n",
    "ridge = linear_model.Ridge()\n",
    "\n",
    "for split in splits :\n",
    "    print(\"Cross validation with %i time splits\" % split)\n",
    "    errors_splits, alphas_splits = time_cross_validation(ridge, X, y, 5, split)\n",
    "    \n",
    "    errors_array = np.array(errors_splits)\n",
    "    \n",
    "    mae_split.append(errors_array[:,0].mean())\n",
    "    rmse_split.append(errors_array[:,1].mean())\n",
    "    r2_split.append(errors_array[:,2].mean())\n",
    "    \n",
    "    print(\"Mean RMSE %.2f\" % errors_array[:,1].mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAp0AAAFFCAYAAACuUFbhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3XeUVFW2x/HvbnISAyYkKkoQFUmi4yhiADHrgAEHFRQjigkVRYIi5hn1mRABdRDFgJExo5hGghEEEZVkwoyIgsB+f5xqu+hQdauprqru/n3WqkXXveeeu5s3M29zwj7m7oiIiIiIlKW8bAcgIiIiIhWfkk4RERERKXNKOkVERESkzCnpFBEREZEyp6RTRERERMqckk4RERERKXNKOkVERESkzCnpFBEREZEyp6RTRERERMpc1WwHUNby8vK8Vq1a2Q5DREREJKlVq1a5u1fIQcEKn3TWqlWL3377LdthiIiIiCRlZr9nO4ayUiEzaRERERHJLUo6RURERKTMKekUERERkTKnpFNEREREypySThEREREpc0o6RURERKTMKekUERERkTKnpLO0rr+ely6fRrNmkJcHzZrBS5dPg+uvz3ZkIiIiIjlHSWcpvfRLJ3a7pjfNF0/DHZovnsZu1/TmpV86ZTs0ERERkZxj7p7tGMpUnTp1vCxOJGrWDJovfoUpHM0zHEp3nqc3k/mi6X4sWpT214mIiEglYGar3L1OtuMoCxrpLKUlS+BV9mM1NTiRidzJGbzKfixZku3IREREpKIys3FmttzM5sRd62Vmc81svZl1TPDspmb2qJnNN7N5ZrZnZqIOMpp0lvAXtZuZvW1mH5nZ02a2SQnPnh/7C51jZpPMrGbmIi+qSRPoyqvUZhUA53IbXZlGkybZjEpEREQquAlAj0LX5gBHA9OTPHsL8Jy7twJ2A+alPboEMj3SOYGif1FjgUvdfRdgCnBx4YfMbDvgXKCju7cFqgDHlW2oiY3tM43J9OYfPMp3NGAubZhMb8b2mZbNsERERKQCc/fpwI+Frs1z908SPRcb1NsHuDf2zBp3/7nMAi1GRpPO4v6igJYUZOYvAseU8HhVoJaZVQVqA1+VSZARHVB/Jh8MmcwnTbtzF2eyJ/9jwq43c0D9mdkMS0RERKQ42wPfAePN7D0zG2tmGV07mgtrOucAh8d+7gU0LtzA3b8EbgSWAF8Dv7j7CyV1aGYDzGyWmc1au3ZtGYQMDB7MAaPCpqGhX5/F+ryq1Jozk8+OGVw27xMREZHKoGp+DhP7DEhXv0B74E533x34Dbg0TX1HkgtJZz/gbDObDdQD1hRuYGabAUcAzYGGQB0zO7GkDt19jLt3dPeOVatWLaOw42yzDWuOPp6T14/jqgszOlItIiIiFcva/Bwm9hmTpn6XAcvc/Z3Y90cJSWjGZD3pdPf57n6Qu3cAJgGfFdPsAOALd//O3f8EHgf2ymScydS+7Dzq8hubPzmOt97KdjQiIiIiBdz9G2CpmbWMXdof+DiTMWQ96TSzrWJ/5gFXAHcV02wJ0MXMapuZEf6iMrrjKqn27Vn3t304v8qtDL5gLRW8/KmIiIhkgZlNAt4GWprZMjPrb2ZHmdkyYE/gWTN7Pta2oZlNjXt8IDDRzD4E2gHXZDT2TBaHj/1FdQUaAN8Cw4C6wNmxJo8Dl7m7m1lDYKy794w9OwI4FlgLvAec6u6rk72zrIrDF2vKFDj6aI7hUY59+Bh6987Ma0VERKRiqMjF4XUiUTqtW4e3aMF7yxvxj61fZ948qFEjM68WERGR8q8iJ51Zn16vUKpUwc49l/ar3mDzL2Zx223ZDkhEREQkN2ikM91WrIBGjXil3hEc/dsDfPYZbLFF5l4vIiIi5ZdGOiW6TTaBfv3ouvxh6q74ipEjsx2QiIiISPYp6SwLAweSt24td7e7kzvugAULsh2QiIiISHYp6SwLO+wAhx/OwYvvYtMav3PJJdkOSERERCS7lHSWlUGDyPvxe8YfMJEnnoDp05M/IiIiIlJRaSNRWXGH3Xdn/Zq1NPnlI7ZtaLzzDuQpzRcREZESaCORpM4sjHbOm8v4E19m1iyYNCnbQYmIiIhkh0Y6y9Lq1dCkCd6xEx2+foYffoD586FWreyEIyIiIrlNI51SOjVqwFlnYVOf5Y5BC1iyBG65JdtBiYiIiGSeRjrL2rffQpMmcNppHL7k/3j1VVi4ELbaKnshiYiISG7SSGcJzNg0XYFUWFtvDSecAOPHc+PlP7FqFYwYke2gRERERDIrUtJpxplmDI773s6MZcAPZsw2o1GZRVgRnHcerFrFTtPHcvrpcPfdMG9etoMSERERyZyoI50DgRVx328FvgL6xPq4Ns1xVSzt2kHXrnDbbQy/Yi116sDgwUmfEhEREakwoiadTYBPAMzYEvgbMNidh4CrgG5lE14FMmgQLF3Klm9MYcgQeOYZeOWVbAclIiIikhmRNhKZ8QNwgjvPm9EbuBfY1J11ZnQFprpTu2xDLZ2sbyTKt24dtGwJW2/NHy+/ScuWsMUWMGuWCsaLiIhIoI1EMAM424ydgXOB59xZF7u3PWGqXRKpUgXOPRfeeouaH85g9Gh47z144IFsByYiIiJSDLOqmNUodO0gzAZh1j7l7iKOdO4MPElIMJcCB7qzIHbveeAbd05K9eWZkDMjnQC//gqNGsGhh7L+gYl06QJffQULFkDtnBwnFhERkUzKqZFOs8eAX3DvF/t+LvBvYDVQBTga92eidhdppNOdue60ALYEmuUnnDEXARdGfWGlVq8e9O8PkyeT9/WX3HQTfPkl3HxztgMTERERKaILMDXu+8XATbjXAsYCl6fSWdSSSePMaO7OD+4UHhpdAVyfyksrtYEDYf16uP12/v53OOoouPZa+OabbAcmIiIisoEtgJChmO0CNATuit17BGiTSmdR13SeTBjlLE4DyM2p9ZzUvDkccUQo1rlqFdddF45ov/LKbAcmIiIisoFvgWaxn3sAi3H/LPa9FrA+lc5S2Tdd0uLPtsB3qby00hs0CH78Ef7zH3bcEc4+G+69F+bMyXZgIiIiIn95BLgOsxuAS4D74+7tDnyaSmclbiQy4zzgvNjXpoTh1dWFmtUEtgYmuNM/lRdnSk5tJMrnDh07wu+/w9y5/PCj0aIFdOkC//1vtoMTERGRbMmxjURVgSFAJ+B94Crc18TuPQ68hfuNkbtLkHQeCBwEGHAB8CDwdaFma4D5wGT3IglpTsjJpBNCraS+feH55+Ggg7jpJrjoor++ioiISCWUU0lnmkUtmTQMGOvOl2UfUnrlbNK5ejU0awa77w5Tp7J6NbRuDXXrhvqdVapkO0ARERHJtJxKOs3WAXviPqOYex2AGbhHzliilkwaUR4TzpxWowacdVaYT58/nxo1wi72jz6CCROyHZyIiIgIluBeNWBtSp0lmF6fDFzmzmexnxNxd45N5cWZkrMjnQDLl0OTJnDKKXDnnbjD3/4GX3wBn34aRj1FRESk8sj6SKdZEwp2rL8KnAnMK9SqJqFyUQfcW0XtumqCe1sSsliArSh597qU1lZbQZ8+cN99MGoUtvnm3HQT7LUX3HgjDB+e7QBFRESkkjkFGEbI+xy4s4R2vwOnptJxpDWd5VlOj3RCmE/fddcwt37JJQAceyw880wY7WzYMMvxiYiISMYkG+k0s3HAocByd28bu9YLGA60Bjq7+6wSnl0E/AqsA9a6e8diGm1JGGw04EOgT+zPeGuAJbintIlcSWcuOOAA+OQT+PxzqFaNzz8Pm4r69IFx47IdnIiIiGRKhKRzH2AlcH9c0tmaUKj9buCiJElnR3f/PmIwTYGv/yqTtJFKnF4346xUOnLnjo0Pp5IaNAgOOwwefxyOPZbttw+nZd58M5x3Huy2W7YDFBERkVzg7tPNrFmha/MAzBLt+4nIrDbuq2LfvgOqxup1lhTQqhLvFe46wUaiVI42cndysshPuRjpXL8eWraEBg3g7bcB+OknaNEC2reHF16AdPznSERERHJblI1EsaTzmfyRzrjrr5J4pPML4CfCWs273X1MMY0KyiSZrSfZnp4USiaVmLm6p3REpmyMvLwwpDlwIPzvf9ClC5ttFs5jHzQoVFXq2TPbQYqIiEgGVDWz+KRxTLHJYen8zd2/MrOtgBfNbL67Ty/Uph/wWdzPaVuHmdE1nSUsft0NuAuoCywC+rj7imKe3RQYSzjr3YF+7v52sneWi5FOgJUroVEj6NEDHnoIgDVrYOedoXp1+OADqJqo1oCIiIiUe2U50lmo7XBgpadwjOXGijyaaUZ1MwaYMdaMZ2N/nmZG9RTeNwHoUejaWOBSd98FmAJcXMKztwDPeagHtRtFa0aVb3XrwqmnwqOPwtKlQEg2r78ePv4Y7r03y/GJiIhIuWVmdcysXv7PhKPO52Q0hojHYLYGngMaArOB5YTt9O2Bb4Ae7nwc6YWFsnMzWwHUd3c3s8bA8+7eptAzmwAfANt7ikOz5WakE2DRIthhB7j44lBCCXCHffcNm9s//RQ22SS7IYqIiEjZibB7fRLQFWgAfEuoqfkjcBuhxvrPwPvu3t3MGgJj3b2nmW1PGNyDsLzyQXcfVcwLZpLKlLp756hNoyadrwP1gUPdWRJ3vQnwLPCTO/tEemHRpPMt4Dp3f9LMLgBGuHu9Qs+0A8YAHxNGOWcD57l7sdmkmQ0ABgBUr169w+rVKZWRyq5eveDll8NoZ53wn7mZM6FzZxgyBEYV/Y+HiIiIVBA5cCLRBFJLOk+J3HXEpPN34Hh3nijm3lHAg+7UivTCoklnK+BWYAvgKeBcd9+i0DMdgf8RFsC+Y2a3ACvcfWiy95WrkU6AN9+EvfeGO++EM87463KfPqGi0oIF0LhxFuMTERGRMpP1pLMMRV3TuYhwzmZxakLB6Geq3H2+ux/k7h2ASRTsmIq3DFjm7u/Evj9KmNqvePbaCzp2hFtuCaWUYq65Jky1X355FmMTERGRysnMMNtyY4qBRk06LwWuNmOPDd9PF2AkcElpA4ht28fM8oArCDvZN+Du3wBLzaxl7NL+EG0NabljFuokzZ8fCnTGNG0aLj/wAMyencX4REREpPIw60lYCvkHYR/PH5i9hdkhKXcVcXp9JtCUMAW+nIKNRFsBPxBGQv/iTrGLSktY/FoXODvW5HHgstimor8Wv8aebUfY6V4d+Bw4xd1/ShZ7uZteh1ArqVkz2GUXeP75vy7/8ksoGN+2LbzyigrGi4iIVDQ5Nb1udjpwB/AyIUfLz/+OJgwAnoX73ZG7i5h0jk8lRnciLyota+Uy6YSwY+iKK2DOnFCsM+aOO+Dss+HJJ+Hww7MYn4iIiKRdjiWdi4GpuJ9ZzL27gJ64N4ncXSaLw2dDuU06v/sOmjSBvn3h7oJ/RPz5ZxgAdQ/5aLVqWYxRRERE0irHks6VwFG4v1jMvQOBKbjXjdpdqY+6NGMzM9qZUaO0fUgCW24JJ54I998PP/zw1+Vq1eCGG8Iu9jHpOhRLREREpKhpwL4l3NsXKHyEZkJRp9dHADXcuTT2vRvwJFCbsKj0IHfmpvLiTCm3I50Ac+eGBZzXXAOXXfbXZXfYf3/46CNYuBDq189ijCIiIpI2OTbSeSBhP81U4AkK1nQeBRwMnAp8+Vd794SbvKMmnQuBq92ZEPv+HiHZHAGMAn5zJydXGJbrpBPgoINC8rlo0QZz6e++GyorXXwxXHdd9sITERGR9MmxpHN9oSsOWKHvxK457lUSdhcx6VxFOOpyuhmNgcVAF3dmmHEIMN6draL+DplU7pPOqVPhkEPgwQfh+OM3uHXSSfDww6G6UrNm2QlPRERE0ifHks6SptaL5/5awu4iJp3fAqe687QZJwE3u7NF7N7+wJPuRF5ImknlPulcvx5atw5z6O+8s0GdpGXLYMcd4aijQk4qIiIi5VtOJZ1pFnUj0WvApbFRzYsI6znz7QQsTXdgEpOXB+edFw5gf/vtDW41agQXXgiTJsGMGVmKT0RERComs60wax733TAbgNm/MTss5e4ijnRuBzwAdALeB3q783Xs3tvAh+6cnurLM6Hcj3QCrFwZDlw/8ECYPHmDW7/+GkY7d9wRpk9XwXgREZHyLKdGOs2mAgtxPzf2fSQwBFgItABOxX1C1O4ijXS686U73dyp587f8xPOmO7AwKgvlFKoWxdOOw0eewwWL97gVr16MHIkvPEGTJmSpfhERESkImoPvAJAOK78TGAI7q0IG8kHpdJZqet05nNnhTtrNrYfSeKcc8Iw5u23F7nVr184tOiSS8IJmiIiIiJpUJ9w3DlAB2BzYGLs+yuE0c7INjrplAxp0gSOOQbuuSdMt8epWjUUjF+4MByTKSIiIpIGy4A2sZ8PAebjnl+Xsz7wRyqdKeksTwYNgp9/DqcUFdKjR1jyOXIk/PRTFmITERGRimYccD1mjwCDgfizELsA81LpTGevlyfu0KVLSDznzQs72+N8+CG0awfnnw833ZSlGEVERKTUcmojEYBZXwo2ko8jP3E0uwt4G/f7InelpLOcmTQJTjgBnnkmFI0vpH9/eOCBkJPusEMW4hMREZFSy7mkM41SSjrNOBjoCDQmHIu5xIx9gIXufFVGMW6UCpd0/vlnOH6oTRt48cUit7/6KpRPOuSQItWVREREJMdV5KQz0ppOM7Y24x3gaeAkoD/QIHb7FGBo2YQnRVSrFnayv/QSzJlT5HbDhjB4MDzyCLz1VhbiExERESlG1I1EtwF1gVaxT3wJ8peA/dMclyQyYADUqgW33FLs7Ysugm23DacVVfDVEyIiIlJORE06ewBXuLMQKJzGLAO2S2tUktgWW0DfvmHx5nffFbldpw5cfTX8739hxFNEREQk21IpmbSuhOsNgN/TEIuk4txzYfVqGDOm2NsnnQS77gqXXhqaiYiIiGRT1KTzdWCgGVXiruWPePYj/4gkyZw2baB793BCUTHHEFWpEsomffEF3HZbFuITERGR8s9sV8wexuwzzFZj1j52fRRmB6fSVdSk8xJCjaY5wFWEhPM0M6YDewJXpPJSSZNBg+Drr0vcpn7AAXDwwWGq/Ycfim0iIiIiUryQVM4GtgHuB6rF3V0NDEylu0hJpztzCGduzgJOJky1Hw0sBfZwZ0EqL5U0OeggaNUK/vWvEncM3XAD/PprOKlIREREJAWjgQm47wuMKnTvfaBdKp1FXtPpzmfu/NOdhu5Ud2cbd/q482kqL5Q0ysuD886Dd9+FN98stsnOO8Npp4Uz2RfonwYiIiISXSvg4djPhUe3VgCbp9KZzl4v7/r2hc02g3//u8QmI0ZAzZphU5GIiIhIRMuB7Uu4tzOwJJXOIiedZvQ2Y5IZ082YUfiTyksljWrXhtNPhylTYNGiYptsvXVIOKdMgenTMxueiIiIlFsPASMx2zvummO2E2G/z8RUOot6ItG1sRdvT1jHObeYj2TL2WeDGfzf/5XY5PzzoVGjUDB+/foMxiYiIiJpY2bjzGy5mc2Ju9bLzOaa2Xoz65jk+Spm9p6ZPRPhdUMJ+3leo2BU80nCxvIPgWtSij3K2etmLAf+5c7oVDrPBRXu7PWSHH88/Pe/sHQp1KtXbJP77w/1O//zH+jTJ8PxiYiISFLJzl43s32AlcD97t42dq01sB64G7jI3WcleP4CoCOwibsfGjGo/QmnTzYAfgRexv3FaL9RgajT638StsxLrho0CH75BSZMKLHJiSdC+/YwZAj8rnL+IiIi5Y67TyckfvHX5rn7J8meNbNGwCHA2BRf+jLuQ3AfgPulpUk4IXrSeQtwqtkGZ65LLtljD+jSJZzHXsL8eV4e3HgjLFlS4rHtIiIikl1VzWxW3GdAGvv+NzCYMCoanVlNzLbHrE2RTwqqRmnkzvVm3AjMN+M14OciTZxLUnmxlIFBg+C44+DZZ+Gww4ptst9+4dY110C/frDVVhmOUURERBJZ6+4J12WWhpkdCix399lm1jXiQ42AMUD34u4SyihVKeZe8d1FXNPZB7iPkBl/BxQ+d9HdS9xSn1WVZk0nwNq1sP32sOOO8PLLJTabPx/atg2b3m+/PYPxiYiISELJ1nTG2jQDnslf0xl3/VVKWNNpZqOBfwJrgZrAJsDj7n5ighdNI2wivxZYSNH8D9xfSxTrBt1FTDqXAtOBM9z5NWrnuaBSJZ0A118Pl1wCH3wAu+5aYrNzzoG77oKPPoLWrTMYn4iIiJSorJLOQu26xtol3khk9ivQB/enkgYeQdQ1nZsA48pbwlkpnXpqqN2ZZNHmsGFQpw4MHpyhuERERGSjmdkk4G2gpZktM7P+ZnaUmS0D9gSeNbPnY20bmtnUjXjdx0DtjY86iDrSOQ74yp0rNuplZuOA/DUF+dv8dwPuAuoCi4A+7r6ihOerEOpFfRl1m3+lG+kEOOssGDcu7BhKsGjzuutC0fiXX4Zu3TIYn4iIiBQrykhnxph1JpRhOgf34s/bTqW7iEnnsYT5/JeBVyi6kQh3kmbSJdSWmkkY4n3NzPoBzd19aAnPp1xbqlImnfPnhznzESPgyitLbPbHH9CqFWy+OcyaFXa3i4iISPbkWNJZHbgNOJWwnrPojLd75C3JUZPOZFvr3T3a7qXC6xDMbAVQ393dzBoDz7t7kS34sdpS9wGjgAuUdCbRsye8+y4sXgw1apTYbNIkOOGEUN7zpJMyF56IiIgUlWNJ5/1AL+ApSt5INCJydxGTzqbJ2rizONILiyadbwHXufuTsZHMEe5e5EgdM3sUGA3UI8ri15hKm3S+8AJ07w733Qd9+5bYbP36UN7zq69gwYKwHFRERESyI8eSzl+Bi3G/Kx3dRZpQdWdxss9GxNAPONvMZhMSyiJZdHxtqSgdmtmA/KKqa9eu3YjQyrEDD4Q2beDf/4YE/7DIy4Obb4Yvvwx/ioiIiMR8R8GZ6xutxJFOM2q7syr/52Qd5bdN+sIStvnH7u0E/MfdOxe6nnptqZhKO9IJcM89MGAAvPYa7LNPwqbHHAPPPw8LF8I222QoPhEREdlAjo10ngz0BQ7HfeVGd5cg6VwH7OnOjNiazoTz8BuxpnMrd19uZnnABOBVdx+X4PmuaHo9mt9/h8aNQ8L5+OMJm376aRgY7dcP7r47Q/GJiIjIBnIs6XwE6ALUIVQPKnIiJe7HRu0u0TGY/YDP4n5OvvgziVhtqa5Ag1g9qWFAXTM7O9bkcWB8rG1DYKy799zY91ZatWqFY4dGj4bPPw+nFZVgxx3h7LPhtttg4MBwYpGIiIhUag0IG4gAqgFbbkxnkTYSlWeVeqQTwmLNZs1CJplk0eaPP8IOO4SNRf/9b2bCExERkQI5NdKZZpE2EpnxuRm7lXCvrRmfpzcsSZvttoPevWHsWFhRbM39v2y+OQwdCs89Fza/i4iIiKRLKnU6u7gzo5h7nYE33KleBvFttEo/0gkwcyZ07hx2sp93XsKmq1eHtZ116sB770GVSCt1RUREJB2yPtJpdhbwCO7fxX5OzP2OyF0n2Ei0CbBp7Osi4Ejg/ULNagJnAEe50zzqSzNJSWfM3nvD11+HYpxJMslHHikYHO3fP0PxiYiISC4kneuBLrjPiP2ciOMeeXgqUdI5jLDRJ9lQqAEXuvOvqC/NJCWdMY8+Cr16wRNPwBFHJGzqDn/7G3zxRdjVXrduhmIUERGp5LKedJahREnnjsBOhKTyKeAi4JNCzdYAn7inr3BouinpjFm7NuwS2n57mDYtafO334a99oJhw2D48LIPT0RERHIs6TTbB3i32BqdZnWADrhPj9xdxDWd+wKz3dnowqCZpqQzzo03wsUXh8Wa7dolbX7ssfDMM2G0s2HDDMQnIiJSyeVY0rkO2BP3Int6MOsAzEhlej3qMZivlceEUwrp3z/sELrllkjNr702DJBecUUZxyUiIiK5yBLcqwvRTqP8qzPV6axkzjknHI+5eHGk8y4vuiiU93zvPdit2KJZIiIiki5ZH+kMU+pdY9+GA2OBZYVa1QQOAX7Dfa/IXSvprGQWLICWLSMv1vzpJ2jRAtq3D7U7LdG/eURERGSj5EDSeTEwOPZtc2AFsLZQqzXAfOBi3N+N3LWSzkrosMNgxoww2lmzZtLmt94ayns++yz01KGkIiIiZSbrSWc8sy+AI3H/IB3dJV3TaUZNM+4xo0s6Xig5YNAgWL4cHnooUvMzzghns198cVjjKSIiIpWAe/N0JZwQIel05w/gOML8vVQE3bpB27bhhKIII93Vq8N118HHH8O992YgPhEREalwIu1eB14B9ivLQCSDzMJo5wcfwGuvRXrkyCPh73+HK69MeoS7iIiISBFR63QeRNi9NBmYCnxLoZOK3Pm4LALcWFrTWYLff4cmTcLRQ088EemR/CPcL78crr66jOMTERGphHJqTWeaRU06C5+9Gf+QAe5O5OKgmaSkM4GhQ2HUqLCjvUWLSI/06QOPPx4eady4jOMTERGpZCpy0hl1en2/Qp9ucZ/871LenHUWVK0Kt90W+ZFrrgnLQC+/vAzjKgcmToRmzSAvL/w5cWK2IxIREUkjs5qYrcbsyLR1qZJJlVzfvjBlCixbBvXrR3rkssvCaUWzZkGHDmUcXw6aOBEGDIBVcecw1K4NY8aEkWAREZHSyqmRTrOlwJm4P5OW7lJJOs04GOgINAaudmeJGfsAC935Kh0BpZuSziRmz4aOHcOxQ+efH+mRX36BRo3gzz9hzZqwNHTUqNxJuNxh9eqwbPX330NymOznqO1WrYJPPoF164q+t2lTWLQo47+uiIhUIDmWdF4B7AMcgvufG91dxDWdWwNPAR2ARUBzoJM775oxHvjDnTM3NpiyoKQzgn32gaVLYeFCqJJ8ae7EiXDKKSHpzJdspG/dutIleKX9ubQD+DVrQq1a4fepVav4nx97rPhnzWB94dXPIiIiKcixpPNG4ATCXp6XKbqR3HG/JHJ3EZPOycDOwBGEpHMN0DGWdPYBhrmzU9SXZpKSzggefxyOOSb8edRRSZs3axYOMyqsRg3YZZfik8A1a0oXWpUqyZPA4n4uzTM1a4Y1mqX9/fPy4P8J76NVAAAgAElEQVT+D047LSyVFRERSVWOJZ1fJGnhuG8fubuISecK4CR3pphRBfiTgqRzX2CqO7nxF1SIks4I1q0Lu9ebNIlUtzMvr+SRxIMPTm9CWK1amn/XNChuTWfNmmF6/ZNPoFUruOEGOOQQnVUvIiKpyamkM81SGY8pZhUbAA2A39MQi2RLlSowcCBceCG8+y60b5+weZMmxY/0NW0KU6eWUYw5JH8JweWXw5IlBWtaTzgBnnwSBg8Ox9t36wY33QTt2mU3XhERqTjMbBxwKLDc3dvGrvUChgOtgc7uPquY52oC04EahPzvUXcflqm4IXrJpNeBgbFRznz5Y139CCcWSXnWvz/UrRuOxkxi1KgwChmvdu1wvbLo0ydsGlq/PvzZp08Y1TzySJg7F269NRz41L59WP/65ZfZjlhERCqICUCPQtfmAEcTksqSrAa6uftuQDugh5l1Sfo2s10xexizz2IllNrHro/C7OBUAo+adF4CdCL8UlcREs7TzJgO7AlckcpLJQfVrw/9+sFDD8HXXyds2qdP2DTUtGlItJo2VbmgeNWqhYHjhQvD4PGDD8JOO8GwYbByZbajExGR8szdpwM/Fro2z90/SfKcu3v+/xeqFvskXmMZksrZwDbA/bFn8q0GBqYSe6Sk0505hFJJs4CTCVPtRwNLgT3cWZDKSyVHDRwIa9fCnXcmbVrcSJ9saNNNw9rO+fPDdPvIkbDjjjB2bPEll0RERICqZjYr7jMgXR2bWRUzex9YDrzo7u8keWQ0MAH3fYHC85nvE0ZMI4s60ok7C935pzsN3anuzjbu9HHn01ReKDmsRYuQHd15J/zxR7ajqTCaNw8DyG+/HX4+7TTYfXd44YVsRyYiIjlorbt3jPuMSVfH7r7O3dsBjYDOZtY2ySOtgIfzHy90bwWweSrvj5R0mjHCjNapdCzl1KBB8P33YU5Y0qpLF3jzTZg8OUyzd+8edvvPmZPtyEREpDJx95+BVym6NrSw5UBJJZF2Bpak8t6oI52nA3PM+MiMIWbskMpLpBzp2hV23TVsKKrgR6Rmgxn06gXz5sGNN8L//ge77RZKMH3zTbajExGRisrMtjSzTWM/1wIOAOYneewhYCRme8ddc8x2Iuz3mZhKDFGTzobAgcBbwCBggRmzzLjQjCapvFBynFkY7fzoI5g2LdvRVFg1aoRNRgsXhqW048eH9Z5XX71h/U8REZF4ZjYJeBtoaWbLzKy/mR1lZssIm7ufNbPnY20bmll+McNtgWlm9iEwk7CmM9mZ6kMJ+3leo2BU80nCxvIPgWtSij2Vs9cBYmWTDgCOBY4E6gNvu7N3wgezRMXhS+GPP8KW9M6d4emnsx1NpfDpp3DJJTBlCmy3HVxzDZx4YrQTkkREpOLIyeLwZvsD+xNqs/8IvIz7iyl3k2rSWfB+ahB2sN8IbONO8kO7s0BJZykNHw4jRsCCBWEITjJi+vQwAjprVqjxedNNYcWDiIhUDjmZdKZJSuMoZlQz4zAz/kNYXHof8DGQtu38kiPOOAOqVw9VziVj9tkH3nknHLX5/few335w+OGh7JKIiEhGmS3C7DrMdk9Hd1F3r/cwYzwh0XwCaAxcBmznzoHu3JuOYCSHbLMNHH98WGz488/ZjqZSycsLR2rOnw+jR8Orr0LbtnDOOfDdd9mOTkREKpFHCcspZ2G2ALORJC+zVKJI0+tmrAdmEHYxTXbnq9K+MNM0vb4R3nsvzPHeeGOY85WsWL48rHS4+26oUweGDIHzzoOaNbMdmYiIpFtOTq+b7UlIPv9B2JA0j1C/8yHcI9drj5p0NnNnUekije+n2EPqdwPuAuoCi4A+7r6i0HONCccvbQOsB8a4+y1R3qmkcyN17RqOHFq4EKpWzXY0ldq8eTB4MDzzTNjnNXo0HHdcKDggIiIVQ04mnfnMDNibkID2BjbHPXJyEPUYzEXhXTQ04xgzTov92TDFcCdQtBDpWOBSd98FmAJcXMxza4EL3b010AU428zapPhuKY1Bg2DxYnjyyWxHUum1bh2KCbz8Mmy2WZiC79IF3ngj25GJiEglUQdoAjQlVC9ancrDUdd0VjHjDmAx8Ahwd+zPxWbcbhY1eS16SD3QEpge+/lF4Jhinvva3d+N/fwrYVh3uyjvlI102GGw/fbwr39lOxKJ6dYNZs+GCRPgyy/h73+HY44Jg9EiIiJpZVYLs96YPUbY23Mv4UjMfsDWqXQVdff6iFjnQ4BmQK3Yn0Ni14en8tJC5gCHx37uRdikVCIzawbsDpR4SL2ZDTCzWWY2a+3atRsRmlClCpx7bji/cebMbEcjMXl5cNJJoaLVyJHw/PPQpg2cfz78WPifdSIiIqVh9jDwHeHkobrA2cA2uB+O+0TcV6bSXdSksy9whTs3uLPEndWxP28gVKs/OZWXFtKPMF0+G6gHrCmpoZnVBR4DBhVe9xnP3ce4e0d371hV6xA33imnQL16cEukZbQV3/XXFz2tadq0cD3DateGoUNDcfmTTgoVrlq0CAPTa0r8b5KIiEgkWwMXAdvi3h338YRz20slatK5FeG4o+J8GLtfKu4+390PcvcOwCTgs+LamVk1QsI50d0fL+37pBQ22QT694eHH4avyk3hgrLTqRP07l2QeE6bFr536pS1kLbdFu65B95/P4RxwQVh5POxx6CU5z+IiEhl594V97tw/z4d3UVNOhcAx5Vw7zjgk9IGYGZbxf7MA64g7GQv3MYIawjmufvNpX2XbISBA2HdOrjjjmxHkj3r14dNVWvWhCSzZ0/o2zf8PHlyqOSeZbvsEqban3sOatWCf/wjrPl8p8TFKCIiIgmYbYrZJZg9jdmbsT8HY7Zpyl1FLJnUm1Cj8xVCodBvCaObvYD9gOPceSR5PzYJ6Eo4u/NbYBgFawQAHgcuc3c3s4bAWHfvaWZ7A68DHxFKJgEMcfepJKGSSWl01FHw+uuwdGnIaCqqP/4I89Xz5oUK7fPnh58/+QR+/72gXV5eSET//vcw2lklt06CXbs21PYfOhS+/TaUVxo9Gpo1y3ZkIiJSkpwqmWS2A/AasCXwJiF32xrYi7CpaD/ci52hLra7qGevm3EQYUNRe6Aa8CcwGxjmTsqHvmeKks40eu21ULfznnvg1FOzHc3G++GHDRPL/OTyiy8K5qTNQlHM1q2hVauCzw8/wGmnwXbbwYcfQrt2MHVqmOfOMb/+CjfcEGr8r18fCssPGQL162c7MhERKSzHks6ngOZAD9y/jLu+HfBf4Avcj4jcXdSks+A95BFGKr93/2vUMWcp6Uwj93BC0Z9/wkcflY+q5OvWhSnxwonl/PnhcPN8NWvCTjttmFy2bg077hh268TLX8OZP6V+ySVhE9Gmm8Ijj8ABB2T2d4xo2TK44gq4/37YYgsYPhwGDIBq1bIdmYiI5MuxpHMFcBLuU4q5dwwwHvdNIneXatJZ3ijpTLP77w/bpF94AQ48MNvRFFi1KtQPKpxYLlgQpsvzNWhQNLFs1QqaNIk+PX799WG3TvwazvHj4dJLw+HoQ4fClVfm3HR7vvfeC6eaTpsGLVuGX+eww8rHvyFERCq6HEs6fwH64/5oMfd6AWNxjzxvpqRTUrN6dZhu7tABnn02s+92D0ld4cRy/vwwmhk/Jd68edHksmXLkHSWld9+g7PPhvvuC8sQHnwwJ6fbIfxVPfssXHxx+Ovr2hVuuikMZIuISPbkWNL5BOEQnx64L4673hR4DpiH+9GRu1PSKSkbORKGDQvZSsuW6e9/7dpw3nvhxHL+/A0rn9eqteE6y/wkc8cdw3R5tkyYAGedFWqbTpyYs9PtEFZK3HNP+D/n99/DP/8Jo0ZB44RHNIiISFnJsaSzGWETeSPgXQo2kncAlgL7474ocndKOiVly5eHrOTUU+H220vfz2+/hR3hhRPLBQs2rGy+1VbFT4k3bhx2kOeiuXPDus9588JCymHDcna6HeCXX+Daa0NRebMw/X7JJSFvFhGRzMmppBPArDrhIJ9OwLbA14RTISfgntIxJEo6JXX5J/JMnx52p2y2Wfg+cyYMHrxhW/dQr6e4XeJLlxa0y8uDHXYomli2bAmbb57Z3y9dfvsNzjknjHzm+HR7vsWLw872Bx+ErbcOg9r9+oEO9hIRyYycSzrTqMSk04yGwHJ3Eh5ebkZdYFd33iqD+Daaks4yMG0aHHMM/PRTSEA7dgyjejfdFOrwFF5z+csvBc/WqVP8lHiLFlCjRvZ+p7IUP93+n//k1gasEsyYEUY733gDdt45lFvq0SPbUYmIVHyVNelcB+zpzozY9zxgIXCYO3Pj2u0BvOVOTs4dKuksI9OmQffuYYRybezfJevWFdzfdtuiiWWrVtCoUeXcJl3OptshDFJPmRKm2RcuhIMOCvU+P/oILr8cliwJm/5HjYI+fbIdrYhIxVBZk871QJe4pLMKoSB8R3fejWunpLOyOumkUEJpp53CaUXxU+Kbpnw6VsVXDqfbISyvvfNOGDEiDG5XqbLhvy9q14YxY5R4ioikQ0VOOnN0F4bkvGnTwgk8Q4eGHeXdu4ckdI89lHCWpE6dUM9zwoQwf92uHbyYs4d5/aV69XCK0WefhRUC8QknhBKpl1+endhERKT8UNIpqYs/kWfkyPBn797huiR30klh01WDBiFZv/LKoplcDtpsM1i5svh7S5ZkNhYRESl/kiWdxc29V+zt7pLczJkFR0BC+HPy5HBdomnTJox2nnwyXHUV7L8/fPVVtqNKqkmT4q/XqxfODRARkXLObAhm2xa6tg9mdQpda47ZmJS6TrKm8xUgvxq3AccALwM/xTXdHNhPazpFSum++8Lu9jp1QjH5HN7dPnFiOK991aqCa/lrPHfeGcaNg86dsxefiEh5l/U1nWbrgD1xnxH7XgVYA3TC/d24dnsAb+EeOf9LNNI5HagCbBn7NABeA6rGXdsy1mZ69N9GRDaQP92+5ZZhun3o0IKKADmmT5+waahp01CEoGnTkDNPnRoqY+25ZyjV+vvv2Y5URERKqbgSM2kpO6Pi8CK54rffYODAsNlo333D7vaGDbMdVWS//BLOcr/nnlDAYNw42GuvbEclIlK+5MBI53qgS6GRzj+BjmU50ikimVSnTsjUJkwII5/t2sELL2Q7qsjq1w+joC++CH/8AXvvDRdcsOFUvIiIVF4lJp1mNDRjz2KutzPjMTPmmvGKGUeVbYgilUz8dHuPHjk93V6cAw4IBeTPPDOc5b7bbuHEVBER2XhmNs7MlpvZnLhrvcxsrpmtN7OOJTzX2Mymmdm8WNvzErymJma1MasN1ClyLVyvlWrsiUY6RwE3bRgwOwKvAwcA84DNgEfN2D/VF4tIAvG726++OmRy5WB3e7569eD220MVrfXrw2qBgQNLLrkkIiKRTQAKH0w8BziaxHts1gIXuntroAtwtpm1KaHtNODX2Cd/8/jrcdd+JWw2T0nVBPf+Btxa6NoFQA3CqUQfApjxBHApYVe7iKRL/nR7165h2LBdu3B2+0EHZTuyyLp2hQ8/DMXjb70Vnn0Wxo6Fbt2yHZmISPnk7tPNrFmha/MALMEx0+7+NfB17OdfzWwesB3wcaGmp6Qx3A0kKpm0EjjCvSCZNONLYKE7+8ZdOxy4y52c3PGgjURSIXz8MfTqFc5uv/zycHZ71UT/Zsw9b7wB/frBp5/C6afD9dfDJptkOyoRkdxiZmuAj+IujXH3MYXaNAOecfe2ha6/Clzk7rOSvKMZYVS0rbuv2OigI0o0vb6KuPl6M5oD21J06PYnQOceipSlcj7dDmFj0QcfwEUXhR3ubduWq31SIiKZstbdO8Z9UirAnoyZ1QUeAwaVOuE02wGzy4lbVxpFoqTzfeCfcd/7EE4jeqZQux2IDdeKSBnKn26/775yubsdoFYtuOEGePPN8Ot07w79+8PPP2c7MhGRis/MqhESzonu/niKD2+L2SDM3gEWAMOAxal0kSjpHAkcbcYHZrwIjACmufNOoXbHQJFrIlJW+vYNSedWW4Xd7VdcUa52twN06QLvvQeXXhpy6LZtw3pPEREpGxYWfN4LzHP3myM+tBlmp2H2CrAUuBnoCNwAbIf7IanEUGLS6c4bwH6EdQWrCbvZNyiPZMaWwHrCTioRyZT86fZTToFRo8rN2e3xataE0aPhf/+DzTaDQw8N+fSPPyZ/VkSksjKzScDbQEszW2Zm/c3sKDNbBuwJPGtmz8faNjSzqbFH/0aYwe5mZu/HPj2LeUEdzE7A7GngG+BuYCvCyGYnwulEU3H/LuXYdSKRSDl3//1hd3udOuVud3u+1atD7jx6NDRoAHfdBUccke2oREQyLwdOJFpJ2NPzBfAwMAn3ObF79Ql7ebrinnIFZp1IJFLe9e0Ls2aV6+n2GjVg5MiwamDrreHII+H44+H777MdmYhIpWMUnLXusU96Ok5QMimlop/u5GTlPY10SqWxalWowD5uHOyzD0yaVK7Obs/3559w7bVw1VWw6aZwxx3wj39kOyoRkczIgZHO2sARwPFAd0JN93nAJOB5YAalHOlMlHSuJ1ScfxVYk6wjd3ql+vJMUNIplU4FmG6HcJTmKafA7Nkh6bz99jCYKyJSkWU96YxntinQCzgO2JeCEdCbgZtw/yal7hIknfcCRxKm4J8gZLgvubO+dJFnh5JOqZTmzQvF5D/+GIYMgeHDy10xeQirBG68MdTCr1cPbrsNjjsOEhy6ISJSruVU0hnPbBtC8nkc0BlYB7yE+8GRu0i0kciMaoSh1eOAw4HfCfWdJrnzeukjzxwlnVJprVoF554L995brqfbIeTO/frBO++EDUZ33gnbbpvtqERE0i9nk854Zs2BE4DjcN8l8mNRd6+bUZOQeB4LHAz8CIxy587Uo80cJZ1S6T3wAJxxRphuf+CBUJG9HFq3Dv7977BPqlat8PM//6lRTxGpWMpF0llKKZdMMmNn4HTgLOApd44ui8DSRUmnCGG6vXdvmDMnTLePGFEup9sBFiwIo55vvgk9e8Ldd0OjRtmOSkQkPbKedJr1TaG14/5A5K6jJJ1m7EDBPH4r4A1C7aZH3cnpoiZKOkVi4qfb//73MN2+3XbZjqpU1q2D//s/uOwyqFYNbr45JKIa9RSR8i4Hks71FJRJSva/qo57lahdl1in04ztzLjAjBmEMzYPIxyf1NSd/dy5K9WE08zGmdlyizsg3sx2M7O3zewjM3vazDYp4dkeZvaJmS00s0tTea+IALVrw9ixYXf77Nnh7Pbnn892VKVSpQqcd17Y4b777nDqqaFE6ZIl2Y5MRKTc+y32eZCQ+9UH6pXwKTZnK0mi3etrYy99GngI+DxRR+58nPRlZvsAK4H73b1t7NpM4CJ3f83M+gHN3X1ooeeqEBLfA4FlwEzgeHdP+k6NdIoUowJNt69fH04wGjw4jHTeeCMMGKBRTxEpn3JgpLMmcChhdrsnYQ/Pw8BDuM/cqK6T1OnMl2gO3gB3J9Lwqpk1A56JSzpXAPXd3c2sMfC8u7cp9MyewHB37x77fhnhpaOTvU9Jp0gJKtB0O8CiRXDaafDSS9CtWxjUbd4821GJiKQm60lnPLO6wNGETeQHEAb+HgIewH1+qt0lOgZzv7hPtwSf/PulNYewKx5CAdLGxbTZDlga931Z7FqxzGyAmc0ys1lry9lxgCIZkz/d/sAD8O675Xq6HaBZM3jhBRgzJhynucsuYd3n+nJVWVhEJIe4r8T9ftwPAbYBHgcGA9eUpruUd69vrGJGOlsBtwJbAE8B57r7FoWe6QV0d/dTY9//CXR294HJ3qeRTpEI5s8PxeQrwHQ7wNKlYYr9uedCidJ774UWLbIdlYhIcjk10glg1ogw0nkc0IFwDOZNuD+SaleJRjojxkILs9LX6nT3+e5+kLt3IJx69FkxzZax4QhoI+Cr0r5TRApp1SpUXu/fH665JsxPf/lltqMqtcaNYepUGD8ePvgAdt0V/vWvsOtdRESSMNsSs7Mwex1YDPyTcDjQ9rh3KU3CCUmSTjOqmtHFjF5mdCx0r5MZjwLzgd6leXnox7aK/ZkHXAHcVUyzmcCOZtbczKoTsu2nSvtOESlGcdPtzz2X7ahKzQxOPjmcZrT//nDBBWHp6vyUVyGJiFQiZs8DXwIDgZeBNri3w/1a3BdtVNcJNhI1BZ4FWhPbLAT8l3Ds0d2ERPMH4F/A/7nza/LfwyYBXYEGwLfAMKAucHasyePAZbFNRQ2Bse7eM/ZsT+DfQBVgnLuPivILanpdpBTip9svuwxGjizX0+3uMHFi2De1alX4dS64oFz/SiJSQWV9ej3U6VwJfELijeSBe+fIXSdIOh8kJIjnAx8ATYHRhLpMTYGRwE3u/B71ZdmgpFOklFatCsUwx46FvfeGhx4q17vbAb75Bs46C6ZMgU6dwvT7zjtnOyoRkQI5kHROIEqymc/9lMhdJ0g6lwJD3ZkQd21X4H3gYnduihxQFinpFNlI//lPOLu9Vq0w9d6jR7Yj2ijuMHkynHMOrFgBV14ZanxWq5btyEREciDpLEOJ1nQ2JKzXjJdfjP2NsglHRHLOiSfCrFmwzTZw8MHQpw/ElyKbNg2uvz578aXIDI49Nqz1PPJIuOIK2GMP+PDDbEcmIlKxJUo6DSi81zN/WHRN2YQjIjkpf3d7z57w4IPh7Mlly0LC2bt3mKsuZ7bcEh5+GB57LGzU79ABhg+HNfpfNxGRMpHsRKJ5hKMw43UkjHiuir/oTuSFpJmk6XWRNBsyBEbHDgOrUgUOOgiOOAI6d4a2bcvlPPUPP4TlqxMnhvJK48dD+/bZjkpEKqOKPL2eKOkcn0pH7kReSJpJSjpFysA558Dtt8MOO8DPP4esDaBmzZCtdeoUktDOnUObcnIQ+lNPheWry5fDpZfC0KFQo0a2oxKRyqRSJp0VhZJOkTTLn1I/80y4884wR928OcyYUfCZPRt+jxW22GyzggQ0Pxndeuvs/g4J/PRTKKc0YULY2T5uXAhZRCQTlHSWY0o6RdIoP+GcPBn226/o93xr18LcuQVJ6MyZ8NFHBQehN2lSkIh27hwWVNatm53fqQT//W84SvOrr+DCC8PJoLVqZTsqEanolHSWY0o6RdLo+uvDaGV8gjltWkgqBw9O/Oxvv8F77204IvrFF+FeXh60br1hIrrLLllfH/rLL3DRRaFUacuWYdTziy/g8sthyZKQO48aFTb0i4ikQ84lnWb/AI4mHEFes8j9dBSHryiUdIrksO++C+WY4hPR778P92rWDLvk46flW7TIyvrQF1+EU08NiWbVqhtWjKpdG8aMUeIpIumRU0mn2XDgSsIhQR9TXPWidBSHryiUdIqUI+6waNGG0/KzZ4fTkSCsD43fpNSpU6gfmgG//goNG8LKlUXvNW0awhYR2Vg5lnQuBR7AfUhaulPSKSI5be3aUMk9fjR0zhxYFysj3KTJholohw5Qr16ZhJKXF/LiwswKlquKiGyMHEs6fwaOwf3ltHSnpFNEyp1Vq4quD/3883DPDNq02XA0dJddoHr1jX5ts2aweHHR602aFH9dRCRVOZZ03gX8ivvFaekuatJpRsKFpCoOLyJZ9f33YTp+5syCRPS778K9GjUK1ofmf0qxPnTixLCjPX+2P9+ee8LLL2t3u4hsvGRJp5mNAw4Flrt729i1XsBwoDXQ2d1nRX02STC9gOuA14AXgZ+LtHGfmrSf/O6iJJ1mDCfJQlIVhxeRnOIehh8L1w/Nzxg33XTDafnOnSOtD504sWD3euPG0KULPPIIdOwITzwR1n2KiJRWhKRzH2AlcH9c0tkaWA/cDVyUIOks8mySYJItHHLcqyTtJ7+7iEnnUuABd9KykDSTlHSKyF/WroV58zZMRD/6qGB9aOPGG+6W79ABNtmk4PkSSkbNnTCTPR4bzKabhlONdISmiJRWlOl1M2sGPFM4cTSzV0mQdCZ6toTGTZO2cY+8uKhqxHb1gLQsIhURyZqqVcP6zl12gf79w7VVq+D99zdMRB97LNwz27B+aP36xRbH33nyZN68AA4/HPbeG+6/H/7xj+z9miJSrlU1s/ikcYy7j8lKJCkklFFEHem8C/jVnbQsJM0kjXSKSMp++KFgbejMmfDOOwXrQ6tVC1P3nTqFU5cmTYKePQH49ls4+mh4661wgtHQoeXm2HkRyRE5NdJZ8FBVoAnFF4f/OGo3UUc6XwauM6MBJSwkdSfyQlIRkZy2xRbQo0f4QEgylywpGAl96CF4++1w74gjwihot25svd9+vPLsngw4rxbDhoVKT+PHa4ORiJRTZtWAW4GTgBoltEr7ms6kC0ndo780kzTSKSJplX/efP/+cNddcMgh8Nln4WSldeugRg18zz15s9p+XPZiN9a278xjT1fXBiMRiSSnRjrNrgJOBgYDE4Gzgd+AE4EdgIFlsXs96UJSd3KySp2SThFJm/yEs9CaTiZPDpuOXn89XHvllbBO1J3fqM3MGnvT4tT9aNS3W9hlVDXqJJOIVDYRdq9PAroCDYBvgWHAj8BtwJaE2ej33b27mTUExrp7z5Kedfd7EwTzCXA9MAH4E+iE++zYvfuAP3A/PfLvpuLwIiIRlbB7nZkzYfDgDdv++CO89hrfPzKNHx59hZZ/zg3X69WDffaBbt1CP7vtFo46EhEh54rDrwK64/567OfDcX8pdu8g4EHcG0TuLpWk04wSF5K6E3khaSYp6RSRbFu+HPodupzaM19lcIdX6PDrNGzBgnBz881h330LktA2bbT7SKQSy7Gk8zNgEO5PYzYXeAz3K2P3zgSuSnvSaUbShaRa0ykiUrLVq+H00+G++/6/vTsPk6q69j7+XdCMgiNDRGnaCcSJFgEVEAEVsWOcZRBFvVe9YoxTHBLNdYjXV6+XOCQ3MUFEIQShPaEAACAASURBVCKKioq8gCCCOEYZRAYFEhlEhkYRBIHuBtb9Y52yiurq7uqh6lR3rc/z1NM1nDq1T4H4673X3ttG5J976BuafBwMxc+cCStX2oGtW0Pv3tEQWoWdk5xztVeGhc5ngO9QvRORW4DhwHigCBgIjEP135M+XZKhs8JC0kydve6h0zmXKVRh+HC46y4r7XzjDTjkkODFFSssfEZqQteutecPPdTCZ58+FkTbVbxWs3Ou9sqw0PkzoAWqi4LHtwKXAE2w1Yx+j2rSISvZ0FmqkFSVucFro4GdqiRdSJpOHjqdc5nmzTfhssusvPONN6xMdC+qsHx5tBd05szoOqGHHRbtBe3Tx/fddK6OyajQWcOSDZ3bgbNVeS+4f54qbwev9QNeUCXpMf108tDpnMtECxfaDkbr18Nzz8HAgeUcrGoL0Ud6QWfNgs3BcskdOkRDaO/e0LJl6hvvnEuZjAydIscAJwFtgVGorkfkSGADqluTPk2SofNfwC2qvCnCYuBVVe4NXhsGPOih0znnKmfjRtvB6P334d574b77kpzIvns3LFgQDaGzZ8O2bfba8cdHQ+jpp8P++6f0GpxzNSujQqdIM2AUcDGwC9tUqCuq8xAZD6xG9fakT5dk6HwG+E6VO0VIWEiqStKFpOnkodM5l8mKimDYMNu56JJLbKJR06aVPElJCcydGw2hH3wAO3bYBKTOnaP1oD172pi+cy5jZVjoHAEUAFcAHwA7gS5B6LwKuJ1KbKeZbOj8GdBClUXB41KFpKpkZLLz0Omcy3Sq8NhjcMcdcOKJVud56KHVOGFRke0XH6kH/egjKC6G+vVty85IPWj37lVIuM65VMqw0PktcDOqYxGpj83riYTOPsBEVJP+TdYXh3fOuQwxaRIMHmydka+/bvmwRmzfbsEzMjHpk09siL5hQzjllOhw/MknQ6Oytld2zqVDhoXOH4GLUZ2aIHSeB4xBNekansouDr9XIakq60U4EtigStKFpOnkodM5V5ssWgS/+IVNMHr2WRg0KAUfsnWrFZJGQui8edbd2qQJ9OgRDaFdukS37KzMbkzOuSrLsNA5C1iL6mUJQucYbDmlgqRPl+TwesJCUlXmiTAeWK1K0oWk6eSh0zlX22zcCBdfbFu5/+538MADKd4p8/vvbTJSpCZ04UJ7vnlzOO00C6HNmlljEu07HxtEnXPVkmGhsyfwNvA+8DLwF2yv9w5YmWUvVD9N+nRJhs6EhaRB6LwKuF2VpAtJ08lDp3OuNioutglGo0ZZAB09GvZJ1/+GNm60ZZkiIXTpUnu+WTNrWM+eMGcOPPGELTjqQ/LO1ZiMCp0AIj2AR4BTgPqAAh8Dd6L6QaVOlWTo/Ba4WZWxIvzUvRqEzj7ARFWSKiQVkVHAuUChBjOeRCQf+Cu2p/su4AZV/STBex8Ffg7UwyYw3awVXICHTudcbaUKjz8Ot98O+fkwcWI1JxhV1dq10UlJr74aXSMUrAu2XTtbL7R9++jP9u2tsSntonWu7sm40Bkh0gQ4ANiM6vaqnCInyeOaAN+V8VpzYHclPvM54H+BMTHPPQo8oKpTRKQgeNw79k0i0h3oAZwQPPU+cDowqxKf7ZxztYYI3HYbHH201XZ27WoTjE4+Oc0NadMGhgyxn2+8YTWcI0bADTdAgwbWE7psmdWJRtYLBasRPeqo0oG0QwdfP9S52kZ1B7CjOqdINnR+CgwFpiZ47RLgw2Q/UFVni0he/NPAvsH9/YC1id6K9YQ2BARoAGxI9nOdc662Kiiwyee/+IWt9z5qlI1qp1V8DWf//tHH999vx6jCunXRELpsmd3/7DOYMMFmzEe0bFk6iLZvD0cc4cP1zoVJ5N5KHK2oPpj0qZMcXq+wkFSV5AtJLXROihle7wi8hYXJekB3VV2V4H3DgWuC4/5XVe8p4/zXAdcBNGzY8KSioqJkm+accxnr22+tvnP2bLjnHvj979M4el3d2evFxbBiRTSQxgbT9eujx9WrB3l5iQPpIYf4cL2r80IfXhfZg/Vo/ojlrfIoqq2SPnWySyaJUGYhqSqVKyQtHTr/CLyrqq+KyADgOlU9M+49RwJPYjsggdV03qWqs8v7LK/pdM7VJcXFNqr9zDO2heaYMWmcYJQqW7bA8uWJA2nsv99Nm+49XB8bSH243tURGRA6/wnkAjOAF4HXUP2hRk5d2cXhRfipkFSVKhWSJgidW4D9VVVFRIAtqrpv3HvuABpr0I0r1v27U1UfLe+zPHQ65+oaVXjySfj1r+GEE2yCUdu2YbcqBVRtElNsEI38XLFi7+H6Vq1KT2Tq0AEOP9yH612tEnrotEZ0AQYBA4AWWHnlOGBSUNtZtdOGsSNRgtD5BTBMVWeJyBnAo6p6Utx7BgLXAv2x7t6pwBOq+mZ5n+Wh0zlXV02ZYhOMmjSx+T1pn2AUpuJi+Oqr0vWjy5bBhphy/3r14LDDyh6ul4pGD51Lr4wInbFEemEB9GKgKTAR+BsVjDQnPFVZoVOEShWSqpJUIamIjMNmprfAJgLdByzFhs5zsDVAb1DVuWJJ+3pVvUZsJfy/AL2wof2pqnpbRZ/nodM5V5ctWWITjL75xobchwwJu0UZYPPmsofrt8cM0DVtWnqYPvJzv/2ix/luTC6NMi50Rog0BB4CbsX2XL+o0qcoJ3RWqpBUlaQLSdPJQ6dzrq779lu45BJ49124+2548EGfb5PQnj3lD9fv2RM9tnXraCCtXx9efBH++EcYPBg++MB3Y3Ipk3Gh0xaHH4RNHG8OTAKeQvXdSp+qnNBZqpBUlRopJE0nD53OuWxQXAw33ghPPw0XXmgTjJo1C7tVtUhRkQ3XJwqkhYXR4+rVsxB67bVw8802sSkLhujHjrUVE1avhtxceOgh71VPlYwInSKdsaA5EGiNlTS+iPVwVmk+D1RQ0ylCmYWkqtVbIDRdPHQ657KFqnXG3XYbHH+8TTDKzQ27VXXA5s0WPh9+2Fbnb9HCupfB1hUtKIBzzoHeva3Ato4ZOxauu650ZcKIER48UyH00CmyFDgMeAcLmhPSPntdhISFpKpUupA0nTx0OueyzdSpMHCg5Z/XXoNTTw27RXVAZHH8YcPgqads+YDNm20214wZsGMHNG4MfftGQ+jhh4fd6mrbtct+cVm3rvRr7drBypVpb1KdlwGhcw82v+ZHbA5N+VKxTme0LexVSKpKpQtJ08lDp3MuG33xhU0w+vprm2B0+eVht6gWi9+NKf7xzp1WUDtlCkyebJOYwCYlFRTY7bTTMn7ppu3b4fPPYf786G3hQqs8SETEXmvQIL3trOsyIHTeV6njVR9I+tSVXBy+VCGpKpUuJE0nD53OuWz13Xc2wWjWLPjNb6wOzycYVUFlZ68vXx4NoLNmWTLbZx8444xoL2jIdQ+bNu0dLufPtxLWyFyqAw6AE0+023PP2d+lRFq1si1Zhw6F/PysKG9NudBDZwpVVNNZZiFpVReGTzcPnc65bFZcDL/6ldXfnX8+PP+8TzBKq+3bLaBOnmy3yHj0scdGe0F79EhZd6GqLacVHzBXxWw0feih0YAZueXmRgNkWTWd119vl/Pmm1BSAscdZ+FzyBBo0yYll5MVsjJ0ilCqkNRnrzvnXO2jCn/6E9x6qwWDiROtHs+lmap1J0YC6OzZltaaN4ezzor2glYxse3ZY52s8QEzMudJxCbbd+4cDZf5+dCyZcXnLm/2+qZN8NJLtmLCxx9bb/qZZ8KVV8IFF1hAdcnL1tBZqUJSX6fTOecy21tv2QSjRo1sglH37mG3KMtt3QrvvBMNoWvW2POdOkV7QU85BXJySr21uBgWL4Z586LhcsGC6Fb1DRrYLxixvZedOqW+l3vZMvj73+22apV93qWXWg9or15e3pGMbA2dlSokVSXpQtJ08tDpnHNRX35pE4xWr7Y1PYcODbtFDrBe0MWLowH0gw9s6vj++1PStx8rji5gdtP+fPiv1syfb4eWlNhbmzWzHsvYgHnMMdCwYXiXs2cPvPee9X6+/LLl69xcuOIKu3XoEF7bMl1FoVNERgHnAoUx24lfCtwPdAS6qeqcMt7bH9sBsj4wUlUfqeHmlyuUvdfTyUOnc87tbdMmm2A0cybcdZcNldavH3arHNg69PPnw5KPtrBn+tvkLppMzx8mczDrAZif04UlueewpUcBLQu6kn9SfY44IrN7ELdvhzfesAA6bZoF0pNPtl94Bg6Egw4Ku4WZJYnQ2QvYBoyJCZ0dgT3A34DbE4XOYDvxZcBZwBrgU2Cwqi6p+asoo+0eOp1zLvuUlMBNN8Ff/wrnnWcTjJo3D7tV2UPVJuHE11+uXRs9Ji8v6LnMV07f7zM6rZvCvu9PRj76yJLbQQdB//5WB3r22bZofYZbtw5eeAFGj7blmBo0gHPPtQBaUBBu72ymSGZ4XUTygEmR0Bnz/CzKDp2nAver6tnB498CqOrDNdPyinnodM65LKUKf/6z7eZ47LE2C9knGNW8Xbts/lB8wNy82V6vVw86dtx7eDw/35YtSmjTJpg+3Ybhp0yBjRttltDJJ1sALSiw2UKZ3P2J1aCOGWOTlDZssAw9aJAF0K5ds3f5JREpBhbGPDVCVUfEHZNH5UPnJUB/Vb0meHwFcLKq3lijF1AOD53OOZflpk2ztc4bNrQJRj16hN2izJPs3uM7dlgPXmy4/PxzWz8ebNOi44+3YBmZRX788dXYPXPPHpg7N7ou6Cef2G8TrVpZAD3nHOjXr5wEG75duyxDjxlju4zu3Gk1n0OH2qYG2baVawp7Oi8Fzo4Lnd1U9Vc11PQKeeh0zjm31wSjESNsuRtnylqn8okn4Mgj9w6YX34Ju3fbMfvtV3r9y6OPTjgZveZs3GjLFEyZYvuhbtpkBbunnhpdkqlTp4ztRtyyBV55xQLo7NnWzN69LYBefHF2lID48Hot5qHTOeeSs2mTLW/zzjtwxx3w8MPZPcGoqMgm9nTrBuvXl39smzalA2ZeXsjZbvdu6/mMDMPPnRttbGQY/swzYd99Q2xk2VassFrjMWPgn/+03uCLLrIAesYZdffvZgpDZw42kegM4BtsItFlqrq4ZlpeMQ+dzjnnflJSYjWeTz1lPZ9jx9ad3iVVq6MsLLQawsLC8u9v2VLxOadOtYDZKiNXqo6zfr01ePJkq6nYssW6XXv2jPaCHntsxvWCqsJHH1n4fOkl+zNs08aG3ocOtSbXJUnMXh8H9AZaABuA+4BNwJ+AlsBm4DNVPVtE2mBLIxUE7y0AnsCWTBqlqg+l8lpKtd1Dp3POuXiRCUYdO9oEo7y8sFuUWHGxjSgnEyILC6NrW8YSsUksrVpB69b2M/b+PffYZ8Rr1y66q2Wts2uXJblIL+iCBfZ827bRhen79rVFQCu793wK7dwJkyZZAJ0yxS6jc2cLn4MH15LwX4GsXBy+rvDQ6ZxzVTN9uk0wysmxCUY9e6b+M1Xhhx+S643csCE6Azxeo0YWGhOFyPj7LVqUX2dZVk3niBGJJxPVSmvWRHtBp0+HbdtsZlmvXrZ35osvWrFl374WOAcMgPHj9w6iaVZYaM0aM8YqB+rXt87aoUOtl75x49CaVi0eOmsxD53OOVd1y5bZOoorV8LVV9sclYpmcMcrKbH9v5PtjSwqSnyeAw9MLkS2bm0ddDU5Spzs7PU6objYdkSK7I60JFg7vF49aN/e9re8+morrMzNtVvLlqEOyy9ebFtvPv88fPONTeIaONACaPfuGVcxUC4PnbWYh07nnKue77+H006z/7HHatzYdjTKzy8/UG7alPi8DRsmHyJbtLCFxF0IVq2ysezHH7ffQnJybFw7VuPGNjSfm2t1B5EwGrm1bZuWrsfdu60jdswYePVV650+4ojo9puHH57yJlSbh85azEOnc85VX7t21stXkQMOSC5Etmplk6ZrUw9UVosMqQ8bZrPMRo6M/qWI3Fatit5ft85qJWK1alU6jMbeWrWq0b8Q27bBhAkWQN95x5rTs6f1fg4YYL2hmchDZy3modM556qvXr3SGQIsI8yda3mhZUvfxrBOiq/hTKams7jYxrljQ2l8QI3/f3OjRtHe0thbpOe0bdsqr6L/9ddWIjF6tK2l2qgRXHCBBdB+/VK8dmoleeisxTx0Oudc9eXlWU6IV6tncLvkpGL2emT9qrIC6erVthF9fEZp2bLi3tJytv9UtV+SxoyxPeC/+87eMmSIBdBMWDffQ2ct5qHTOeeqLytmcLvMUlJScW/ptm17v6dhw8S9pbG1pftYnisutgn7o0fbsmAlJbYl6dChcNllthYojz7K21u6cs3YPj9NIhs5ZCZn7pe65aI8dNZiHjqdc65mZNUMbpf5VG2B+/JC6dq1tj99rIMOKjXZaeuBuUz7IpdnZ+QyeV5rpF49zjoL+jeeyZA3BjCA8cyiD72ZyXgGsODu8Zz5UGqWi/LQWYt56HTOOeeyVEmJBc/ygunWrXu9RRs04Pt92rJ0Ry7LinIR9nAhrzOR8+jHNAYwnhXt+qSsrMRDZy3modM555xzZSqjt1RXr2b1+6s5hG/IYTcAv+d33MeDiJTuQK0pHjprMQ+dzjnnnKuKvDw4ctXbvMRAJnARF/C693RWQ9lTvJxzzjnnstjIITMZx2Au4RWu42kGMJ7xDGDkkJlhN61W8tDpnHPOOZfAmft9yoK7rWdTBFa062OTiPb7NOym1Uo+vO6cc845lyF8eN0555xzzrlq8NDpnHPOOedSLq2hU0RGiUihiCyKeS5fRD4Wkc9EZI6IdCvjvbkiMk1EvhCRJSKSl652O+ecc8656kl3T+dzQP+45x4FHlDVfODe4HEiY4D/UdWOQDegMFWNdM4555xzNSutoVNVZwOb4p8G9g3u7wesjX+fiBwD5Kjq9OA821R1e/xxzjnnnHMuM+WE3QDgFuAtERmOheDuCY5pD2wWkQnAYcDbwG9UdXf6mumcc84556oqEyYSDQNuVdW2wK3AMwmOyQFOA24HugKHA1eVdUIRuS6oD52za9eumm+xc84555yrlEwInVcCE4L7L2P1mvHWAPNV9StV3QW8DnQu64SqOkJVu6hql5ycTOjMdc4555zLbpmQyNYCpwOzgL7A8gTHfAocICItVXVjcNycZE6+fft2FZEdNdTWsuQA2dqlms3XDtl9/dl87ZDd1+/Xnr2y+frTde1N0vAZoUjrjkQiMg7oDbQANgD3AUuBJ7E/zJ3ADao6V0S6ANer6jXBe88C/gAIMBe4TlWL09b4cojIHFXtEnY7wpDN1w7Zff3ZfO2Q3dfv156d1w7Zff3ZfO01Ja09nao6uIyXTkpw7BzgmpjH04ETUtQ055xzzjmXQplQ0+mcc8455+o4D501Y0TYDQhRNl87ZPf1Z/O1Q3Zfv1979srm68/ma68Raa3pdM4555xz2cl7Op1zzjnnXMp56KwGEWkrIjNF5AsRWSwiN4fdpnQRkcYi8omILAiu/YGw25RuIlJfROaLyKSw25JuIrJSRBaKyGciktTyZXWFiOwvIq+IyJfBf/unht2mdBGRDsGfeeT2g4jcEna70kVEbg3+vVskIuNEpHHYbUoXEbk5uO7F2fBnLiKjRKRQRBbFPHegiEwXkeXBzwPCbGNt5KGzenYBv1bVjsApwC+DfeKzQRHQV1U7AflAfxE5JeQ2pdvNwBdhNyJEfVQ1PwuXEHkSmKqqRwOdyKK/A6q6NPgzz8dWHdkOvBZys9JCRA4BbgK6qOpxQH1gULitSg8ROQ64Ftu8pRNwrogcFW6rUu45oH/cc78BZqjqUcCM4LGrBA+d1aCq61R1XnB/K/Y/n0PCbVV6qNkWPGwQ3LKmQFhEDgV+DowMuy0ufURkX6AXwXa9qlqsqpvDbVVozgD+paqrwm5IGuUATUQkB2iKbW6SDToCH6vq9mBXwHeBC0NuU0qp6mxgU9zT5wOjg/ujgQvS2qg6wENnDRGRPOBE4B/htiR9guHlz4BCYLqqZs21A08AdwJ7wm5ISBSYJiJzReS6sBuTRocDG4Fng9KKkSKyT9iNCskgYFzYjUgXVf0GGA6sBtYBW1R1WritSptFQC8ROUhEmgIFQNuQ2xSG1qq6DqzTCWgVcntqHQ+dNUBEmgGvAreo6g9htyddVHV3MMx2KNAtGIKp80TkXKBQVeeG3ZYQ9VDVzsA5WFlJr7AblCY5QGfgKVU9EfiRLBxiE5GGwHnAy2G3JV2C+r3zgcOANsA+InJ5uK1KD1X9AvhvYDowFVhA9m6F6arBQ2c1iUgDLHCOVdUJYbcnDMHw4ixK17/UVT2A80RkJfAi0FdEng+3SemlqmuDn4VYTV+3cFuUNmuANTG9+q9gITTbnAPMU9UNYTckjc4EVqjqRlUtASYA3UNuU9qo6jOq2llVe2HDzsvDblMINojIwQDBz8KQ21PreOisBhERrLbrC1V9LOz2pJOItBSR/YP7TbB/kL8Mt1Xpoaq/VdVDVTUPG2J8R1WzoscDQET2EZHmkftAP2z4rc5T1fXA1yLSIXjqDGBJiE0Ky2CyaGg9sBo4RUSaBv/2n0EWTSITkVbBz1zgIrLvzx9gInBlcP9K4I0Q21IrpXXv9TqoB3AFsDCobQS4W1Unh9imdDkYGC0i9bFfXsaratYtHZSlWgOv2f93yQFeUNWp4TYprX4FjA2GmL8Crg65PWkV1PSdBfxH2G1JJ1X9h4i8AszDhpbnk1071LwqIgcBJcAvVfX7sBuUSiIyDugNtBCRNcB9wCPAeBH5d+yXkEvDa2Ht5DsSOeecc865lPPhdeecc845l3IeOp1zzjnnXMp56HTOOeeccynnodM555xzzqWch07nnHPOOZdyHjqdy0Iicr+IqIi8leC1V0RkVhrb0jtoS0buaCUiHUXkPRH5MWhnXoJjGgbfaX7c83nBe85NV3tTRURmBUsGRR7fLyLfxjxuHzy3fzgtdM5lOg+dzmW3fiLSNexGZLj/AfbHtn08Fdt3O15DbB2//Ljn1wXveT+VDQzJSODsmMftse/AQ6dzLiFfHN657LUJ29bxHuCCkNuSMiLSWFV3VuMURwMTVXVGZd+oqkXAx9X47Iylqmuwvz/OOZcU7+l0Lnsp8P+wfeSPL+ug+GHUmOdVRG6MebxSRIaLyG9EZJ2IbBGRP4gpEJHFIrJVRF4XkQMSfFQbEZkUDGOvFpHrE3xmTxF5V0S2i8h3IvJ0ZEvO4PWrgnZ1C4aDdwB3lHNt+SIyIzjf9yIyVkRaB6/liYgCRwC3BuedVcaptgY/nw2O0+D9pYbXq/M9iciBIvI3EdkgIjtF5EMRObms6wve0yD4vNUiUiQia0XktWBHpdjvrGtQRrBDRJaJyIUVnPenvxci0ht4M3hpRXC+lcFr+4vIyOBzdwbteLq8czvn6iYPnc5lt5eBZVhvZ00YBHTDtoZ8FLgNeAx4EPhP4HrgdODhBO99Bvgc29d5CvBUXFjrAcwA1gOXALcABcCzCc41DpgUvJ5we1YRaQnMApoCl2HbW54OTA8CWWRofD3wQnD/hjKuu2/w87+C48oaho+o9PckIo2At7EtKO/Aeqc3Am+LyM/K+azfAkOC856FfW9bgPpxx72E7SV9EbAQeFlEOpVz3ljzgNuD+xdh1x8JrY8BPYFbseH4u7FfeJxzWcaH153LYqq6R0QeAZ4RkXtVdVk1T7kTuFRVdwNTReR8LMwdpaorAIIgcyUWrGJNUdW7g/tvicjhwO+IhsZHgA9VdWDkDSLyDTBDRI5T1UUx5/qjqj5ZQVt/Hfw8W1V/CM63DPgHcLGqjgM+FpEiYJ2qljdM/mnw81+xx4ntT59IVb6ny4HjgGNVdXlwzNvA0uBayurR7Qa8oKqjY54bn+C4kao6PDjvW8ASLLAOKusiIlT1BxFZGjycr6or4z7/z6r6Usxzz1d0Tudc3eM9nc6554HVWMCorllBkIr4J7AyEqRinmsZGd6N8Vrc4wnASSJSX0SaYr1n40UkJ3LDJuiUACfFvff/J9HWbsC0SOAEUNVPgJVYz1wqVeV7OhOYiw1fR64f4F2gSzmf9RlwlYjcKSInSNlJ+KfvX1X3YL2e3ZK/pHI//w4RuUFE2tfA+ZxztZSHTueynKruwoZ4LxeRdtU83ea4x8VlPCfYjO9YhQke5wAtgAOw4eC/YCEzcisCGgBt4967IYm2HlzGcRuAA5N4f3VU5XtqAZzC3tdfgg3Rx19/rP8C/oyVBiwAvhaRmxMcl+j7P7jcq0jOjcDrwL3AUhFZLiIV9p465+oeH153zgGMwoay70rw2k7iAmIZE4Gqq1WCx7uAb4HGWB3g/cDkBO9dG/c4mZrBdQk+E6A11qOYaTYBc4BhCV4rKutNwcz9e4F7ReQobLj+CRFZqqpTYw5tBXwX97i8utSkqOpm4CbgJhE5AbgTGCsin6vqkuqe3zlXe3hPp3MusrTPcODfKN27tQZoLiKHxDzXLwXNiJ8tfSEwV1V3q+qP2NJDHVR1ToJbfOhMxj+As+Nmv3cF8qj8uprFwc/GVWhHsmYARwKrE1z/wmROENSC3o6F1GPiXv7p+xeResD5wCeVaF+F34Gqfo7VntbDlqJyzmUR7+l0zkX8DZtZ3B2rE4yYCuwARonIH4DDKD0JqCacIyIPBZ99ETbT+vyY1+/EJg3tAV7BlinKBX4O3FOFSVCPYb2Gb4nIfwPNsMlKC4FXK3MiVS0WkRXAABFZhPUOf17J9lRkDPa9zxKR4cBXwEFY3eV6VX080ZtE5DWs53Y+9ud4CfZv/+y4Q68RkWJgEXAtFnAHV6J9kYlE/yEiLwLbVXWhiLyP1YsuwnqgrwV+pHKB1jlXB3hPp3MOAFXdDpQKLqr6LXAxcChWm3c5tsRQTbsG6Bx8xrnAL1V1Ykw73gd6AS2Bv2PrQt4Js1rgWQAAAOZJREFUfE1yNZx7UdWNQB8sII7D6h7fA85S1eLy3luG67G6y7ex2extqnCOMgXD5H2A6cADwDTgSeAoyg9wH2LLK72ATQ46CZudPyfuuEFYb+frQCdgoKrOr0T7VmG9qBcBHxBdt/Mj4CrsF4Xx2Hd0TrC4vHMui4iqL5fmnHPZSkSuwtY6ba6q20JujnOuDvOeTuecc845l3IeOp1zzjnnXMr58LpzzjnnnEs57+l0zjnnnHMp56HTOeecc86lnIdO55xzzjmXch46nXPOOedcynnodM4555xzKeeh0znnnHPOpdz/AfKfWtq84yirAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax1 = plt.subplots(figsize=(10,5))\n",
    "ax1.plot(splits, rmse_split, 'bo-')\n",
    "\n",
    "ax2 = ax1.twinx()\n",
    "ax2.plot(splits, mae_split, 'rx-')\n",
    "ax1.set_xlabel(\"Number of time splits\", fontsize=15)\n",
    "ax1.set_ylabel(\"Mean RMSE over time splits\", color='b', fontsize=15)\n",
    "ax2.set_ylabel(\"Mean MAE over time splits\", color='r', fontsize=15);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The CV displays a best result for 10 time splits, AND the model seems better than the main Ridge Regression with all airports and all airlines"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "BUT, this is only a regression model toward departure time of the day"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 8.3) Same data but ONE airline, ALL airports"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1) Loading dataframe (pickle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1265267 entries for this airport\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TAIL_NUM</th>\n",
       "      <th>ORIGIN_AIRPORT_ID</th>\n",
       "      <th>ORIGIN_CITY_NAME</th>\n",
       "      <th>ARR_DELAY_NEW</th>\n",
       "      <th>DISTANCE</th>\n",
       "      <th>TIME_SECONDS</th>\n",
       "      <th>ARR_TIME_SECS</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>338103</th>\n",
       "      <td>N551WN</td>\n",
       "      <td>12391</td>\n",
       "      <td>Islip, NY</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1052.0</td>\n",
       "      <td>19800</td>\n",
       "      <td>31200.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>337880</th>\n",
       "      <td>N210WN</td>\n",
       "      <td>11697</td>\n",
       "      <td>Fort Lauderdale, FL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>581.0</td>\n",
       "      <td>19800</td>\n",
       "      <td>27600.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       TAIL_NUM  ORIGIN_AIRPORT_ID     ORIGIN_CITY_NAME  ARR_DELAY_NEW  \\\n",
       "338103   N551WN              12391            Islip, NY            0.0   \n",
       "337880   N210WN              11697  Fort Lauderdale, FL            0.0   \n",
       "\n",
       "        DISTANCE  TIME_SECONDS  ARR_TIME_SECS  \n",
       "338103    1052.0         19800        31200.0  \n",
       "337880     581.0         19800        27600.0  "
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open(path + '/data/data_2016_1_airline.pkl', 'rb') as f :\n",
    "    data_1_airline = pickle.load(f)\n",
    "print(\"%i entries for this airport\" % data_1_airline.shape[0])\n",
    "data_1_airline.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2) One Hot encoding departure airports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [],
   "source": [
    "oh_dep_airport = pd.get_dummies(data_1_airline.ORIGIN_AIRPORT_ID)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "3) Standardise numerical features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [],
   "source": [
    "std_scale = preprocessing.StandardScaler().fit(data_1_airline.loc[:, ['TIME_SECONDS', 'ARR_TIME_SECS', 'DISTANCE']].values)\n",
    "data_num_1_airline = pd.DataFrame(std_scale.transform(data_1_airline.loc[:, ['TIME_SECONDS', 'ARR_TIME_SECS', 'DISTANCE']].values), columns=['DEP_TIME_SECS_STD', 'ARR_TIME_SECS_STD', 'DISTANCE_STD'])\n",
    "# Reindex\n",
    "data_num_1_airline.index = oh_dep_airport.index\n",
    "# Concat\n",
    "data_1_airline_reg = pd.concat((oh_dep_airport, data_num_1_airline), axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "4) Concatenate everything into processed data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>10140</th>\n",
       "      <th>10257</th>\n",
       "      <th>10279</th>\n",
       "      <th>10397</th>\n",
       "      <th>10423</th>\n",
       "      <th>10529</th>\n",
       "      <th>10599</th>\n",
       "      <th>10693</th>\n",
       "      <th>10713</th>\n",
       "      <th>10721</th>\n",
       "      <th>...</th>\n",
       "      <th>14893</th>\n",
       "      <th>14908</th>\n",
       "      <th>15016</th>\n",
       "      <th>15304</th>\n",
       "      <th>15370</th>\n",
       "      <th>15376</th>\n",
       "      <th>DEP_TIME_SECS_STD</th>\n",
       "      <th>ARR_TIME_SECS_STD</th>\n",
       "      <th>DISTANCE_STD</th>\n",
       "      <th>ARR_DELAY_NEW</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>338103</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-1.666111</td>\n",
       "      <td>-1.292933</td>\n",
       "      <td>0.654746</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>337880</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>-1.666111</td>\n",
       "      <td>-1.485725</td>\n",
       "      <td>-0.367602</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 91 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        10140  10257  10279  10397  10423  10529  10599  10693  10713  10721  \\\n",
       "338103      0      0      0      0      0      0      0      0      0      0   \n",
       "337880      0      0      0      0      0      0      0      0      0      0   \n",
       "\n",
       "            ...        14893  14908  15016  15304  15370  15376  \\\n",
       "338103      ...            0      0      0      0      0      0   \n",
       "337880      ...            0      0      0      0      0      0   \n",
       "\n",
       "        DEP_TIME_SECS_STD  ARR_TIME_SECS_STD  DISTANCE_STD  ARR_DELAY_NEW  \n",
       "338103          -1.666111          -1.292933      0.654746            0.0  \n",
       "337880          -1.666111          -1.485725     -0.367602            0.0  \n",
       "\n",
       "[2 rows x 91 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data_1_airline_reg = pd.concat((data_1_airline_reg, data_1_airline.loc[:, 'ARR_DELAY_NEW']), axis=1)\n",
    "display(data_1_airline_reg.head(2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Target : arrival delay\n",
    "y = data_1_airline_reg.loc[:, 'ARR_DELAY_NEW']\n",
    "X = data_1_airline_reg.loc[:, data_1_airline_reg.columns  != 'ARR_DELAY_NEW']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "7) Cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cross validation with 2 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 22.19, MAE : 12.49, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 17.78, MAE : 11.31, R² : -0.00\n",
      "Mean RMSE 19.99\n",
      "Cross validation with 3 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 20.31, MAE : 11.40, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 21.04, MAE : 12.36, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 18.06, MAE : 11.34, R² : -0.00\n",
      "Mean RMSE 19.80\n",
      "Cross validation with 4 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 19.09, MAE : 10.43, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 24.00, MAE : 13.62, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 17.27, MAE : 11.07, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 18.74, MAE : 11.57, R² : 0.00\n",
      "Mean RMSE 19.77\n",
      "Cross validation with 5 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 18.90, MAE : 10.19, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 21.33, MAE : 12.05, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 22.87, MAE : 13.35, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 16.55, MAE : 10.74, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 18.93, MAE : 11.66, R² : -0.00\n",
      "Mean RMSE 19.72\n",
      "Cross validation with 6 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 18.20, MAE : 9.98, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 19.00, MAE : 10.80, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 24.94, MAE : 14.18, R² : -0.05\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 18.77, MAE : 11.70, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 16.72, MAE : 10.72, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 19.33, MAE : 11.84, R² : -0.00\n",
      "Mean RMSE 19.49\n",
      "Cross validation with 7 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 17.80, MAE : 9.86, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 17.95, MAE : 10.27, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 22.37, MAE : 12.67, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 24.10, MAE : 14.02, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 17.52, MAE : 11.15, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 16.08, MAE : 10.44, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 19.86, MAE : 12.10, R² : -0.00\n",
      "Mean RMSE 19.38\n",
      "Cross validation with 8 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 17.09, MAE : 9.41, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 18.89, MAE : 10.49, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 19.95, MAE : 11.33, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 24.85, MAE : 14.17, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 21.21, MAE : 12.82, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 16.17, MAE : 10.57, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 16.66, MAE : 10.61, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 8 : RMSE : 20.23, MAE : 12.29, R² : -0.01\n",
      "Mean RMSE 19.38\n",
      "Cross validation with 9 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 15.88, MAE : 8.96, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 19.21, MAE : 10.49, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 18.87, MAE : 10.75, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 22.75, MAE : 12.95, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 25.03, MAE : 14.57, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 18.19, MAE : 11.46, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 16.28, MAE : 10.57, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 8 : RMSE : 16.34, MAE : 10.44, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 9 : RMSE : 20.88, MAE : 12.60, R² : -0.01\n",
      "Mean RMSE 19.27\n",
      "Cross validation with 10 time splits\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 1 : RMSE : 15.44, MAE : 8.83, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 2 : RMSE : 19.88, MAE : 10.75, R² : -0.03\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 3 : RMSE : 17.67, MAE : 10.24, R² : -0.00\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 4 : RMSE : 20.95, MAE : 11.89, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 5 : RMSE : 22.29, MAE : 12.93, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 6 : RMSE : 25.35, MAE : 14.80, R² : -0.04\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 7 : RMSE : 16.79, MAE : 10.86, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 8 : RMSE : 16.49, MAE : 10.64, R² : -0.01\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 9 : RMSE : 15.92, MAE : 10.31, R² : -0.02\n",
      "[====================================================================================================] 100%\n",
      "Model regression for split number 10 : RMSE : 21.25, MAE : 12.77, R² : -0.01\n",
      "Mean RMSE 19.20\n"
     ]
    }
   ],
   "source": [
    "splits = [2, 3, 4, 5, 6, 7, 8, 9, 10]\n",
    "mae_split = []\n",
    "rmse_split = []\n",
    "r2_split = []\n",
    "\n",
    "# Model\n",
    "ridge = linear_model.Ridge()\n",
    "\n",
    "for split in splits :\n",
    "    print(\"Cross validation with %i time splits\" % split)\n",
    "    errors_splits, alphas_splits = time_cross_validation(ridge, X, y, 5, split)\n",
    "    \n",
    "    errors_array = np.array(errors_splits)\n",
    "    \n",
    "    mae_split.append(errors_array[:,0].mean())\n",
    "    rmse_split.append(errors_array[:,1].mean())\n",
    "    r2_split.append(errors_array[:,2].mean())\n",
    "    \n",
    "    print(\"Mean RMSE %.2f\" % errors_array[:,1].mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax1 = plt.subplots(figsize=(10,5))\n",
    "ax1.plot(splits, rmse_split, 'bo-')\n",
    "\n",
    "ax2 = ax1.twinx()\n",
    "ax2.plot(splits, mae_split, 'rx-')\n",
    "ax1.set_xlabel(\"Number of time splits\", fontsize=15)\n",
    "ax1.set_ylabel(\"Mean RMSE over time splits\", color='b', fontsize=15)\n",
    "ax2.set_ylabel(\"Mean MAE over time splits\", color='r', fontsize=15);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The CV displays a best result for 10 time splits, AND the model seems better than the main Ridge Regression with all airports and all airlines"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 9) Model to keep"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The best model so far seems to be the ONE airline / ONE airport model. And this makes sens.\n",
    "\n",
    "In fact if an airline or an airport is reasponsible for more delays than other airlines or other airports, learning over all airlines or airports won't necessarily help in predicting delays for our specific airline or airport."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Last grid search for our test model :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "80917 entries for this airport and this airline\n"
     ]
    }
   ],
   "source": [
    "with open(path + '/data/data_2016_1_airline_1_aiport.pkl', 'rb') as f :\n",
    "    data_1_airline_1_aiport = pickle.load(f)\n",
    "print(\"%i entries for this airport and this airline\" % data_1_airline_1_aiport.shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Standardise departure time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [],
   "source": [
    "std_scale = preprocessing.StandardScaler().fit(data_1_airline_1_aiport.loc[:, ['TIME_SECONDS', 'ARR_TIME_SECS', 'DISTANCE']].values)\n",
    "data_num_1_airline_1_airport = pd.DataFrame(std_scale.transform(data_1_airline_1_aiport.loc[:, ['TIME_SECONDS', 'ARR_TIME_SECS', 'DISTANCE']].values), columns=['DEP_TIME_SECS_STD', 'ARR_TIME_SECS_STD', 'DISTANCE_STD'])\n",
    "# Reindex\n",
    "data_num_1_airline_1_airport.index = data_1_airline_1_aiport.index\n",
    "# Concat\n",
    "data_1_airline_1_aiport_reg = pd.concat((data_1_airline_1_aiport['ARR_DELAY_NEW'], data_num_1_airline_1_airport), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ARR_DELAY_NEW</th>\n",
       "      <th>DEP_TIME_SECS_STD</th>\n",
       "      <th>ARR_TIME_SECS_STD</th>\n",
       "      <th>DISTANCE_STD</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>339550</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.575404</td>\n",
       "      <td>-1.140637</td>\n",
       "      <td>1.489495</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>339642</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.426590</td>\n",
       "      <td>-0.718927</td>\n",
       "      <td>0.377094</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        ARR_DELAY_NEW  DEP_TIME_SECS_STD  ARR_TIME_SECS_STD  DISTANCE_STD\n",
       "339550            0.0          -1.575404          -1.140637      1.489495\n",
       "339642            0.0          -1.426590          -0.718927      0.377094"
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_1_airline_1_aiport_reg.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Target : arrival delay\n",
    "y = data_1_airline_1_aiport_reg.loc[:, 'ARR_DELAY_NEW']\n",
    "X = data_1_airline_1_aiport_reg.loc[:, data_1_airline_1_aiport_reg.columns  != 'ARR_DELAY_NEW']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From previous test we saw that splitting our dataset in 7 time splits gave the smallest RMSE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_splits = 7\n",
    "# Create time splits\n",
    "tscv = model_selection.TimeSeriesSplit(n_splits=n_splits)\n",
    "\n",
    "# Initialise outputs\n",
    "errors_cv = []\n",
    "\n",
    "best_rmse = 9999\n",
    "# Model\n",
    "model = linear_model.LinearRegression()\n",
    "best_model = copy.deepcopy(model)\n",
    "# Training and testing sets\n",
    "X_train_final = X\n",
    "y_train_final = y\n",
    "X_test_final = X\n",
    "y_test_final = y\n",
    "\n",
    "# Time cross-validation\n",
    "for train_index, test_index in tscv.split(X):\n",
    "    # Create train and test sets\n",
    "    X_train_cv, X_test_cv = X.iloc[train_index, :], X.iloc[test_index, :]\n",
    "    y_train_cv, y_test_cv = y.iloc[train_index], y.iloc[test_index]\n",
    "    \n",
    "    # Training the model with training set\n",
    "    model.fit(X_train_cv, y_train_cv)\n",
    "    \n",
    "    # Prediction\n",
    "    prediction = model.predict(X_test_cv)\n",
    "    # Get Root Mean Squared Error\n",
    "    mse = metrics.mean_squared_error(y_test_cv, prediction)\n",
    "    rmse = math.sqrt(mse)\n",
    "    # Get Mean Absolute Error\n",
    "    mae = metrics.mean_absolute_error(y_test_cv, prediction)\n",
    "    # Get Determination Coefficient\n",
    "    r2 = metrics.r2_score(y_test_cv, prediction)\n",
    "    \n",
    "    if rmse < best_rmse :\n",
    "        best_rmse = rmse\n",
    "        best_model = copy.deepcopy(model)\n",
    "        X_train_final = X_train_cv\n",
    "        y_train_final = y_train_cv\n",
    "        X_test_final = X_test_cv\n",
    "        y_test_final = y_test_cv\n",
    "        \n",
    "    # Save errors\n",
    "    errors_cv.append([rmse, mae, r2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Linear regression on training set : RMSE : 18.62, MAE : 11.11, R² : 0.03\n",
      "Linear regression on testing set : RMSE : 14.34, MAE : 9.49, R² : -0.01\n"
     ]
    }
   ],
   "source": [
    "display_metrics(\"Linear regression on training set\", best_model, X_train_final, y_train_final)\n",
    "display_metrics(\"Linear regression on testing set\", best_model, X_test_final, y_test_final);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 10) Prophet prediction over mean delays by day - WORK TO BE DONE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The best model so far seems to be the ONE airline / ONE airport model. Instead of using linear regression or other kind of regression over flight departure time, let's try to forecast delays based on our time-serie."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Loading prophet module"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [],
   "source": [
    "from fbprophet import Prophet"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Loading data for one airline and one airport"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "80917 entries for this airport and this airline\n"
     ]
    }
   ],
   "source": [
    "# Getting current path\n",
    "path = os.getcwd()\n",
    "with open(path + '/data/data_2016_1_airline_1_aiport_prophet.pkl', 'rb') as f :\n",
    "    data_1_airline_1_aiport_prophet = pickle.load(f)\n",
    "print(\"%i entries for this airport and this airline\" % data_1_airline_1_aiport_prophet.shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Converting date time to date (prophet requirement)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_1_airline_1_aiport_prophet.loc[:, 'ds'] = data_1_airline_1_aiport_prophet.loc[:, 'DATE_TIME'].apply(lambda x : x.date())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looking at our data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TAIL_NUM</th>\n",
       "      <th>ORIGIN_CITY_NAME</th>\n",
       "      <th>ARR_DELAY_NEW</th>\n",
       "      <th>DISTANCE</th>\n",
       "      <th>DATE_TIME</th>\n",
       "      <th>TIME_SECONDS</th>\n",
       "      <th>ARR_TIME_SECS</th>\n",
       "      <th>ds</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>339550</th>\n",
       "      <td>N707SA</td>\n",
       "      <td>Chicago, IL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1521.0</td>\n",
       "      <td>2016-01-01 06:15:00</td>\n",
       "      <td>22500</td>\n",
       "      <td>30240.0</td>\n",
       "      <td>2016-01-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>339642</th>\n",
       "      <td>N8645A</td>\n",
       "      <td>Chicago, IL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>997.0</td>\n",
       "      <td>2016-01-01 07:00:00</td>\n",
       "      <td>25200</td>\n",
       "      <td>39360.0</td>\n",
       "      <td>2016-01-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>339605</th>\n",
       "      <td>N8614M</td>\n",
       "      <td>Chicago, IL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1444.0</td>\n",
       "      <td>2016-01-01 07:25:00</td>\n",
       "      <td>26700</td>\n",
       "      <td>38400.0</td>\n",
       "      <td>2016-01-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>339573</th>\n",
       "      <td>N8318F</td>\n",
       "      <td>Chicago, IL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>990.0</td>\n",
       "      <td>2016-01-01 07:30:00</td>\n",
       "      <td>27000</td>\n",
       "      <td>40560.0</td>\n",
       "      <td>2016-01-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>339512</th>\n",
       "      <td>N8631A</td>\n",
       "      <td>Chicago, IL</td>\n",
       "      <td>0.0</td>\n",
       "      <td>793.0</td>\n",
       "      <td>2016-01-01 07:45:00</td>\n",
       "      <td>27900</td>\n",
       "      <td>37440.0</td>\n",
       "      <td>2016-01-01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       TAIL_NUM ORIGIN_CITY_NAME  ARR_DELAY_NEW  DISTANCE           DATE_TIME  \\\n",
       "339550   N707SA      Chicago, IL            0.0    1521.0 2016-01-01 06:15:00   \n",
       "339642   N8645A      Chicago, IL            0.0     997.0 2016-01-01 07:00:00   \n",
       "339605   N8614M      Chicago, IL            0.0    1444.0 2016-01-01 07:25:00   \n",
       "339573   N8318F      Chicago, IL            0.0     990.0 2016-01-01 07:30:00   \n",
       "339512   N8631A      Chicago, IL            0.0     793.0 2016-01-01 07:45:00   \n",
       "\n",
       "        TIME_SECONDS  ARR_TIME_SECS          ds  \n",
       "339550         22500        30240.0  2016-01-01  \n",
       "339642         25200        39360.0  2016-01-01  \n",
       "339605         26700        38400.0  2016-01-01  \n",
       "339573         27000        40560.0  2016-01-01  \n",
       "339512         27900        37440.0  2016-01-01  "
      ]
     },
     "execution_count": 168,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_1_airline_1_aiport_prophet.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [],
   "source": [
    "max_aircraft = data_1_airline_1_aiport_prophet['TAIL_NUM'].value_counts().index.values[0]\n",
    "data_1_airline_1_aiport_prophet = data_1_airline_1_aiport_prophet[data_1_airline_1_aiport_prophet['TAIL_NUM'] == max_aircraft]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Initialise prophet dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Aggregating delays by day"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ds</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016-01-20</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016-01-24</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016-01-25</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016-01-26</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016-01-27</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           ds    y\n",
       "0  2016-01-20  0.0\n",
       "1  2016-01-24  0.0\n",
       "2  2016-01-25  0.0\n",
       "3  2016-01-26  2.0\n",
       "4  2016-01-27  0.0"
      ]
     },
     "execution_count": 189,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_mean = pd.DataFrame(columns=['ds', 'y'])\n",
    "data_mean.loc[:, 'ds'] = data_1_airline_1_aiport_prophet.loc[:, 'ds'].unique()\n",
    "data_mean.loc[:, 'y'] = data_1_airline_1_aiport_prophet.loc[:, 'ARR_DELAY_NEW'].groupby(data_1_airline_1_aiport_prophet.loc[:, 'ds']).apply(lambda x : x.mean()).values\n",
    "data_mean.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Standardise target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_mean.loc[:, 'y_ori'] = data_mean.loc[:, 'y']\n",
    "std_scale = preprocessing.StandardScaler().fit(data_mean.loc[:, 'y'].values.reshape(-1, 1))\n",
    "data_mean.loc[:, 'y'] = std_scale.transform(data_mean.loc[:, 'y'].values.reshape(-1, 1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train and Test datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_size = 0.3\n",
    "n_train_mean = math.floor(data_mean.shape[0] * (1 - test_size))\n",
    "\n",
    "df_train_mean = data_mean.iloc[:n_train_mean, :]\n",
    "df_test_mean = data_mean.iloc[n_train_mean:, :]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Prophet model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:fbprophet.forecaster:Disabling yearly seasonality. Run prophet with yearly_seasonality=True to override this.\n",
      "INFO:fbprophet.forecaster:Disabling daily seasonality. Run prophet with daily_seasonality=True to override this.\n",
      "C:\\Users\\simon\\Anaconda3\\lib\\site-packages\\pystan\\misc.py:399: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  elif np.issubdtype(np.asarray(v).dtype, float):\n"
     ]
    }
   ],
   "source": [
    "prophet_model = Prophet() #instantiate Prophet\n",
    "prophet_model.fit(df_train_mean); #fit the model with your dataframe"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Prophet future (testing) data to predict over"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "First day : datetime.date(2016, 9, 8), Last day : datetime.date(2016, 12, 28), Number of testing days : 112\n"
     ]
    }
   ],
   "source": [
    "first_test_day = df_test_mean.loc[:,'ds'].iloc[0]\n",
    "last_test_day = df_test_mean.loc[:,'ds'].iloc[-1]\n",
    "delta_days_test = last_test_day - first_test_day\n",
    "print(\"First day : %r, Last day : %r, Number of testing days : %i\" % (first_test_day, last_test_day, delta_days_test.days + 1))\n",
    "date_list_test = [first_test_day + datetime.timedelta(days=x) for x in range(0, delta_days_test.days + 1)]\n",
    "future_data = pd.DataFrame(data=date_list_test, columns=['ds'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Prophet past (training) data to predict over"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "First day : datetime.date(2016, 1, 20), Last day : datetime.date(2016, 9, 7), Number of training days : 231\n"
     ]
    }
   ],
   "source": [
    "first_train_day = df_train_mean.loc[:,'ds'].iloc[0]\n",
    "last_train_day = df_train_mean.loc[:,'ds'].iloc[-1]\n",
    "delta_days_train = last_train_day - first_train_day\n",
    "print(\"First day : %r, Last day : %r, Number of training days : %i\" % (first_train_day, last_train_day, delta_days_train.days))\n",
    "date_list_train = [first_train_day + datetime.timedelta(days=x) for x in range(0, delta_days_train.days)]\n",
    "past_data = pd.DataFrame(data=date_list_train, columns=['ds'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Prophet forecast"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ds</th>\n",
       "      <th>trend</th>\n",
       "      <th>yhat_lower</th>\n",
       "      <th>yhat_upper</th>\n",
       "      <th>trend_lower</th>\n",
       "      <th>trend_upper</th>\n",
       "      <th>additive_terms</th>\n",
       "      <th>additive_terms_lower</th>\n",
       "      <th>additive_terms_upper</th>\n",
       "      <th>weekly</th>\n",
       "      <th>weekly_lower</th>\n",
       "      <th>weekly_upper</th>\n",
       "      <th>multiplicative_terms</th>\n",
       "      <th>multiplicative_terms_lower</th>\n",
       "      <th>multiplicative_terms_upper</th>\n",
       "      <th>yhat</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016-09-08</td>\n",
       "      <td>0.139028</td>\n",
       "      <td>-1.177910</td>\n",
       "      <td>1.231490</td>\n",
       "      <td>0.139028</td>\n",
       "      <td>0.139028</td>\n",
       "      <td>-0.178241</td>\n",
       "      <td>-0.178241</td>\n",
       "      <td>-0.178241</td>\n",
       "      <td>-0.178241</td>\n",
       "      <td>-0.178241</td>\n",
       "      <td>-0.178241</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.039213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016-09-09</td>\n",
       "      <td>0.139781</td>\n",
       "      <td>-1.208161</td>\n",
       "      <td>1.258389</td>\n",
       "      <td>0.139781</td>\n",
       "      <td>0.139781</td>\n",
       "      <td>-0.101532</td>\n",
       "      <td>-0.101532</td>\n",
       "      <td>-0.101532</td>\n",
       "      <td>-0.101532</td>\n",
       "      <td>-0.101532</td>\n",
       "      <td>-0.101532</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.038249</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016-09-10</td>\n",
       "      <td>0.140533</td>\n",
       "      <td>-0.958504</td>\n",
       "      <td>1.450355</td>\n",
       "      <td>0.140533</td>\n",
       "      <td>0.140533</td>\n",
       "      <td>0.071276</td>\n",
       "      <td>0.071276</td>\n",
       "      <td>0.071276</td>\n",
       "      <td>0.071276</td>\n",
       "      <td>0.071276</td>\n",
       "      <td>0.071276</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.211810</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016-09-11</td>\n",
       "      <td>0.141286</td>\n",
       "      <td>-1.099904</td>\n",
       "      <td>1.483127</td>\n",
       "      <td>0.141286</td>\n",
       "      <td>0.141286</td>\n",
       "      <td>0.056894</td>\n",
       "      <td>0.056894</td>\n",
       "      <td>0.056894</td>\n",
       "      <td>0.056894</td>\n",
       "      <td>0.056894</td>\n",
       "      <td>0.056894</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.198180</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016-09-12</td>\n",
       "      <td>0.142038</td>\n",
       "      <td>-1.407673</td>\n",
       "      <td>1.170467</td>\n",
       "      <td>0.142038</td>\n",
       "      <td>0.142038</td>\n",
       "      <td>-0.327769</td>\n",
       "      <td>-0.327769</td>\n",
       "      <td>-0.327769</td>\n",
       "      <td>-0.327769</td>\n",
       "      <td>-0.327769</td>\n",
       "      <td>-0.327769</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.185730</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          ds     trend  yhat_lower  yhat_upper  trend_lower  trend_upper  \\\n",
       "0 2016-09-08  0.139028   -1.177910    1.231490     0.139028     0.139028   \n",
       "1 2016-09-09  0.139781   -1.208161    1.258389     0.139781     0.139781   \n",
       "2 2016-09-10  0.140533   -0.958504    1.450355     0.140533     0.140533   \n",
       "3 2016-09-11  0.141286   -1.099904    1.483127     0.141286     0.141286   \n",
       "4 2016-09-12  0.142038   -1.407673    1.170467     0.142038     0.142038   \n",
       "\n",
       "   additive_terms  additive_terms_lower  additive_terms_upper    weekly  \\\n",
       "0       -0.178241             -0.178241             -0.178241 -0.178241   \n",
       "1       -0.101532             -0.101532             -0.101532 -0.101532   \n",
       "2        0.071276              0.071276              0.071276  0.071276   \n",
       "3        0.056894              0.056894              0.056894  0.056894   \n",
       "4       -0.327769             -0.327769             -0.327769 -0.327769   \n",
       "\n",
       "   weekly_lower  weekly_upper  multiplicative_terms  \\\n",
       "0     -0.178241     -0.178241                   0.0   \n",
       "1     -0.101532     -0.101532                   0.0   \n",
       "2      0.071276      0.071276                   0.0   \n",
       "3      0.056894      0.056894                   0.0   \n",
       "4     -0.327769     -0.327769                   0.0   \n",
       "\n",
       "   multiplicative_terms_lower  multiplicative_terms_upper      yhat  \n",
       "0                         0.0                         0.0 -0.039213  \n",
       "1                         0.0                         0.0  0.038249  \n",
       "2                         0.0                         0.0  0.211810  \n",
       "3                         0.0                         0.0  0.198180  \n",
       "4                         0.0                         0.0 -0.185730  "
      ]
     },
     "execution_count": 195,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forecast_data = prophet_model.predict(future_data)\n",
    "forecast_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ds</th>\n",
       "      <th>yhat</th>\n",
       "      <th>yhat_lower</th>\n",
       "      <th>yhat_upper</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016-09-08</td>\n",
       "      <td>-0.039213</td>\n",
       "      <td>-1.177910</td>\n",
       "      <td>1.231490</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016-09-09</td>\n",
       "      <td>0.038249</td>\n",
       "      <td>-1.208161</td>\n",
       "      <td>1.258389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016-09-10</td>\n",
       "      <td>0.211810</td>\n",
       "      <td>-0.958504</td>\n",
       "      <td>1.450355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016-09-11</td>\n",
       "      <td>0.198180</td>\n",
       "      <td>-1.099904</td>\n",
       "      <td>1.483127</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016-09-12</td>\n",
       "      <td>-0.185730</td>\n",
       "      <td>-1.407673</td>\n",
       "      <td>1.170467</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          ds      yhat  yhat_lower  yhat_upper\n",
       "0 2016-09-08 -0.039213   -1.177910    1.231490\n",
       "1 2016-09-09  0.038249   -1.208161    1.258389\n",
       "2 2016-09-10  0.211810   -0.958504    1.450355\n",
       "3 2016-09-11  0.198180   -1.099904    1.483127\n",
       "4 2016-09-12 -0.185730   -1.407673    1.170467"
      ]
     },
     "execution_count": 196,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forecast_data[['ds', 'yhat', 'yhat_lower', 'yhat_upper']].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plotting results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "prophet_model.plot(forecast_data);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAoAAAAGoCAYAAADW2lTlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3XlAlHX+B/D3MMN9znCIMMBwqYAgCniAR5auSkU/y5JOyxQr29W2dN3tl/VzO6x12611a5cuKzXa7NBSMI8WLUtELS9QUlQOgUEuEQVm5vv7Y2AQFR3UYa7365+Y4XlmPvNxgg/v+T7PIxFCCBARERGR3XAwdwFERERE1Lc4ABIRERHZGQ6ARERERHaGAyARERGRneEASERERGRnOAASERER2RkOgERERER2hgMgERERkZ3hAEhERERkZ2TmLuBG8/Pzg0ql6tU+7e3tcHR0NE1BVoR90GMf9NgHPfahC3uhxz7osQ96ltaH48ePo7a29qrb2dwAqFKpUFhY2Kt9KisrERQUZKKKrAf7oMc+6LEPeuxDF/ZCj33QYx/0LK0PycnJRm3Hj4CJiIiI7AwHQCIiIiI7wwGQiIiIyM5wACQiIiKyMxwAiYiIiEykVaPFpsNqFJY1mLuUbmzuKGAiIiIicyo93YLc4hrkFtdga0ktWtq1uCu+P9Y8bNwRun2BAyARERHRdWjVaLH9WB02FOmHvuKaZgBAsLcLbo0NQJpKgTviAs1cZXd9MgDm5eVh3rx50Gq1mDVrFhYtWtTt+6+//jreffddyGQy+Pv74/3330dYWBh+/vlnPP7442hqaoJUKsWzzz6L6dOn90XJRERERD06XqdP+b7cexI7yn/G2TYtnKQSDAv2xu/HRiBVJccolRyBni5wcJCYu9xLmHwA1Gq1mDt3LjZt2gSlUomUlBRkZGQgNjbWsM3QoUNRWFgINzc3vP3221i4cCE+/fRTuLm54aOPPkJ0dDQqKyuRlJSESZMmwcfHx9RlExERERl0pny5xTXILapBUUfKF+guQ/qgAKSFKzAmXIGBAR5wd7b8D1hNXmFBQQGioqIQEREBAMjMzMTatWu7DYDjx483fD1y5EisXLkSADBgwADD/UFBQQgICIBareYASERERCZ3oiPl21BUg62/1uJsmxaOHSnfU2MjkKaSI9KlFQnRKotM+a7E5ANgRUUFQkJCDLeVSiV27tzZ4/bvvfcepkyZcsn9BQUFaGtrQ2Rk5CXfy87ORnZ2NgCgqqoKlZWVvapRrVb3antbxT7osQ967IMe+9CFvdBjH/RssQ9tWh12ljfju9ImbC1tREndeQD6lO9mlQeGB7ljpNIT4QpXuDpKAQBqdROqqk6Zs+xrYvIBUAhxyX0SyeWn5JUrV6KwsBD5+fnd7j916hQefPBBfPjhh3BwuPTMNVlZWcjKygKgvwbetVyTz5Ku42dO7IMe+6DHPuixD13YCz32Qc8W+nCyvsXwse6Wklo0d6R8Q4O9MX9wf6SpFBgVJkd/r57X8lljH0w+ACqVSpSVlRlul5eXX7ZRmzdvxksvvYT8/Hw4Ozsb7m9qasKtt96KF198ESNHjjR1uURERGTD2jQ6fF9a1/HRbjUOVevX8vX3dMakgf76tXwRCgwK8ISHFazlu1Ymf2UpKSkoKSlBaWkpgoODkZOTg9WrV3fbZu/evZgzZw7y8vIQEBBguL+trQ1Tp07FQw89hLvvvtvUpRIREZENulzKJ3PQr+WbPyYcaeFXT/lsjckHQJlMhuXLl2PSpEnQarWYOXMm4uLisHjxYiQnJyMjIwMLFixAc3OzYcgLDQ3FunXr8J///Afbtm3D6dOnsWLFCgDAihUrkJiYaOqyiYiIyEq1aXT44XjHefmKanCw+gwAILAj5UtVKTA20vZTvivpk1ednp6O9PT0bvctWbLE8PXmzZsvu98DDzyABx54wKS1ERERkfUrqz9nuPrG5iNqQ8o3NNgL88eEI1UlR6pKgUAvF0jtJOW7Evsce4mIiMiqtWt1+KG06+obB6q6Ur7fdKR84yIUGNTPflO+K2FHiIiIyCqUN5xDXnENNhTXYPORWpxp1RhSvnljwpHWcfWN/l6uTPmuggMgERERWaR2rQ47jnelfPtP6VO+fp7OmBDthzSVHGMjfTEowBOeLhxpeoPdIiIiIotR0XgOuUWda/lq0dSR8iUGeeF3o8ORFi5HKlO+68YBkIiIiMymM+XLLVIjt7ga+zpTPg8n3MKUz2TYSSIiIupTl0v5pBekfPojduUI8mbKZyocAImIiMik2rU6/Hi83nD1jc6UL8DDCTdH+yFVJce4CF/E9GPK11fYZSIiIrrhKhvPI6/jvHzfHlYbUr4h/b3w29EqpKkUTPnMiAMgERERXTeNVocfT9Qbjtj9pbIJAODvzpTPEvFfgIiIiK7JhSnfpiNqNJ7vSvmeTNOnfGnhTPksEQdAIiIiMkpnypdbXIN1+ypwUH0OgD7luynSt+Mau76I7ecBLxdHM1dLV8IBkIiIiHp0qql7ytdwTgOpBIj1d8WTaSqkquQYHa5gymdlOAASERGRgUarw08dKV9ucQ32VujX8vm5O2FshC9SVXKMjfCFH5oRrQo1c7V0rTgAEhER2bmqpvPIK1brj9g9UmNI+RKCvDA39cKUzwUyqQMAoLLyvJmrpuvBAZCIiMjOaLQ67DzZYEj59pQ3AtCnfGPCfZEWLseYcF8M7u/JtXw2igMgERGRHag+09rtvHz159ohlQDx/XtO+ch2cQAkIiKyQVqdwM4L1vLt7kj5fN0cMTpcgTSVHKMjFBgc6AVvV6Z89oYDIBERkY2oPtOKjYdrkFtUg2+PqFHX0g4HCZDQ3wtPpIYhVaXA6HAFgpny2T0OgERERFbqSilfapgcaeEKjGHKR5fBAZCIiMiK1HSkfBsuSvni+3vh8dQww9U3lN6uTPmoRxwAiYiILJhWJ1BwsivlKyxjykfXjwMgERGRhelM+XKL1dh4uMaQ8g0O9MJjo8IMB3Aw5aNrxQGQiIjIzLQ6gV1lDcgt6kj5yhsgBCB3dcQolRxpHVffGNzfCz5M+egG4ABIRERkBrXNrdh4WI0NRTXYeLgGpw0pnyfmjAxDqkqOMUz5yEQ4ABIREfUBrU6gsKzr6hu7yi5N+cZE+CKeKR/1AQ6AREREJtKZ8uUW1yCvWJ/ySaBP+bJGhCJVpT+AQ+njCkemfNSHOAASERHdIDqdQGF511q+gotSvtQwfcqXEMSUj8yrTwbAvLw8zJs3D1qtFrNmzcKiRYu6ff/111/Hu+++C5lMBn9/f7z//vsICwsDAHz44Yd48cUXAQD/+7//ixkzZvRFyUREREa5MOXbeFiN2rNtTPnI4pl8ANRqtZg7dy42bdoEpVKJlJQUZGRkIDY21rDN0KFDUVhYCDc3N7z99ttYuHAhPv30U9TV1eH//u//UFhYCIlEgqSkJGRkZEAul5u6bCIiosvSiY7z8l2U8vm4yjAqrOMau+G+SAjyhNzNydzlEl2WyQfAgoICREVFISIiAgCQmZmJtWvXdhsAx48fb/h65MiRWLlyJQBg48aNmDhxIhQKBQBg4sSJyMvLw7333mvqsomIiAxOn23rOC9fDXIPVeP0OQ0kAOICPTF7RCjSmPKRlTH5AFhRUYGQkBDDbaVSiZ07d/a4/XvvvYcpU6b0uG9FRcUl+2RnZyM7OxsAUFVVhcrKyl7VqFare7W9rWIf9NgHPfZBj33oYk+90AmBfdUt+K60EVtLm7D31FkIAN7OUiT6yZCq8sOIYA8M8neDp0vHWr7WRqirG81ad1+yp/fDlVhrH0w+AAohLrlPIpFcdtuVK1eisLAQ+fn5vdo3KysLWVlZAIDk5GQEBQX1us5r2ccWsQ967IMe+6DHPnSx5V6cPtuGby84YlfdsZYvtp8nZo0I7bj6hi8czzcgVBls7nItgi2/H3rDGvtg8gFQqVSirKzMcLu8vPyyjdq8eTNeeukl5Ofnw9nZ2bDvf//732773nTTTaYumYiI7IBOJ7CnolH/sW5RDXaerIdOAN4uMowKkyNVpcDoCDkSg7y7reWrrLSflI9sl8kHwJSUFJSUlKC0tBTBwcHIycnB6tWru22zd+9ezJkzB3l5eQgICDDcP2nSJPzpT39CfX09AODbb7/FK6+8YuqSiYjIRtW1dE/5apr1KV9MPw88Orwr5QuVcy0f2TaTD4AymQzLly/HpEmToNVqMXPmTMTFxWHx4sVITk5GRkYGFixYgObmZtx9990AgNDQUKxbtw4KhQLPPfccUlJSAACLFy82HBBCRER0NTqdwN6OlG9DL1I+IlvXJ+cBTE9PR3p6erf7lixZYvh68+bNPe47c+ZMzJw502S1ERGRbanvlvKpUd3c2i3lS+245BpTPrJnvBIIERFZNZ1O4OfKrrV8P57oSvlGhsk7Bj4FhgR5Q8GUjwgAB0AiIrJC9S1t2HSkFhuKqg0pHwDE9vPAzOEhSA1TGFI+JxlTPqKLcQAkIiKLJ4TAzxVN2FBc3S3l83LunvIlBjPlIzIGB0AiIrJInSlf5xG7VWf0KV9MgAceSQnpuPoGUz6ia8EBkIiILEJnyqc/YrcaP51sgFYn4OUsw4gwHzw2Kgxp4QoMUzLlI7peHACJiMhsGs61Y9MRNXKL9NfZ7Uz5BgV4YEayEqkqOcZG+CJM7saUj+gG4gBIRER9RgiBfaeasKFj4NtxvB5anYCnswwjQ30wZ2QYRkcoMDTYG77uTPmITIUDIBERmVRjZ8rXsZavskmf8g30d8dDSUqkhesP4FDJ3ZnyEfURDoBERHRDdaZ8uUU12HBByufhJMXIMDlmjQjF6HAFhil9mPIRmQkHQCIium6N59qxuUSN3CI1courDSnfgM6UTyXHmEimfESWggMgERH1mhAC+0+dMRyxu+N4PTQXpXxpKgWSQnygcHOERCIxd8lEdAEOgEREZJQzrVp8se8UNhTVIO9wDSoazwPQp3wPJnUcsRupQJjcDc4yqZmrJaIr4QBIRESXJYTAgaozhrV8P5SehkYHeDhJMSJMjpnDQzBapT8vn6+7E1M+IivCAZCIiAyazrdjS0mtPuUrrkF5Z8rn545pMXLcEhuC0eEKhPsy5SOyZhwAiYjs2IUpX25xDb4vrYNGJ+DuJMWIUB88nKIf+JKU3mhtrEVwcLC5SyaiG4ADIBGRnTlzXqM/Yre4BrlFXSlftJ877h8W3HGN3UtTvsomfsRLZCs4ABIR2TghBA5W6Y/Y7Uz52rX6lG94R8qXFi5Hcsd5+biWj8j2cQAkIrJBZ85rsKUz5SuuQVmDPuWL8nPDvYnBSAvXX2OXa/mI7BMHQCIiGyCEwKHq5o4jdqsvSflmJIcgVSVHcogP/JjyEdm9qw6Anp6eV/xB0dTUdEMLIiIi4zS3arClpNawlu9kwzkAXSlfqkqOMZG+iFC4wcWRKR8RdbnqAHjmzBkAwOLFixEYGIgHH3wQQgisWrXK8D0iIjI9IQSKqpuxoeOI3e2lp7tSvhAfPJgUjNRwBVKY8hHRVRj9EfDGjRuxc+dOw+3HH38cI0aMwMKFC01SGBER6VO+rSW12HBRyhfp64bMxOCOa+wy5SOi3jF6AJRKpVi1ahUyMzMhkUjwySefQCrlDxsiohupM+XrPHhj+7HTaNMKuDnq1/I9mBSMUSp9yufvwZSPiK6N0QPg6tWrMW/ePMybNw8SiQRpaWlYvXq1KWsjIrILnSlf59B3ol6f8kX4umF6YhBSVQqMZcpHRDeQ0QOgSqXC2rVrTVkLEZFdEEKguKbZcPDGto6Uz9XRAcND5LhvqP6j3ZRQOVM+IjIJowdAtVqNd955B8ePH4dGozHc//7775ukMCIiW3K2VYOtv3YdsXu8M+VTuOGeIUFIVckxNtIPkb5M+YjI9IweAO+44w6MGTMGEyZM4No/IqKrEELgcE3XWr78o10pX0qID+4dqj9Ny3CmfERkBkYPgC0tLXj11Vev6Uny8vIwb948aLVazJo1C4sWLer2/W3btmH+/PnYt28fcnJyMG3aNMP3Fi5ciPXr10On02HixIl44403+IOSiCzS2VYNvjt6Wn8y5qJqQ8oXrnDF3R0p37gIX0T6uTPlIyKzMnoAvO2227Bhwwakp6f36gm0Wi3mzp2LTZs2QalUIiUlBRkZGYiNjTVsExoaihUrVmDZsmXd9t2xYwd++OEH7Nu3DwAwevRo5Ofn46abbupVDUREpiCEwBH1WcPHuvnHTqNVo4OLzAEpoT7IHBqM1DA5RoQx5SMiy2L0APjGG2/g5ZdfhpOTE5ycnCCEgEQiueqVQAoKChAVFYWIiAgAQGZmJtauXdttAFSpVAAABweHbvtKJBKcP38ebW1tEEKgvb0d/fr1M7ZkIqIbrqVNg+9+PW04GXNpXQsAQCV3xbSE/oaUL8LPHa5M+YjIQhk9AF7rVT8qKioQEhJiuK1UKrudUPpKRo0ahfHjx6N///4QQuDJJ59ETEzMJdtlZ2cjOzsbAFBVVYXKyspe1ahWq3u1va1iH/TYBz32Qa+mpgZH685ja2kjvittwo/lZ9CmFXCRSjA00A23Dw/A8CB3DOnvCbmrTJ/yiWbUq5tRb+7ibzC+J/TYBz32Qc9a+2D0ANh5+bfS0lI899xzKCsrw6lTpzB8+PCr7ncxYz8G+fXXX1FUVITy8nIAwMSJE7Ft2zaMHTu223ZZWVnIysoCACQnJyMoKMiox7/Qtexji9gHPfZBz1770Jny5RbX4JsD1TjRWAagM+XTr+W7KdI+Uz57fU9cjH3QYx/0rLEPRg+ATzzxBBwcHLB161Y899xz8PDwwNy5c7Fr164r7qdUKlFWVma4XV5ebnSjvvzyS4wcORIeHh4AgClTpuCnn366ZAAkIroeQgiU1J5FbsfHuv892rWWL7GfKzKHhSBVpcCIMDkCuJaPiGyA0QPgzp07sWfPHgwdOhQAIJfL0dbWdtX9UlJSUFJSgtLSUgQHByMnJ8foK4iEhobinXfewR//+EcIIZCfn4/58+cbWzIRUY/OtWvx3a+1hqHv6Gn9Wr4wuSvujA9EmkqBsZEKeLSfQXio0szVEhHdWEYPgI6OjtBqtYa/fNVq9SUHbVz2CWQyLF++HJMmTYJWq8XMmTMRFxeHxYsXIzk5GRkZGdi1axemTp2K+vp6fP3113j++edx8OBBTJs2DVu3bkV8fDwkEgkmT56M22+//dpfLRHZtRJ113n5/vvraZzX6OAs05+X766OAzhGhim6pXyVlWfNXDUR0Y1n9AD4u9/9DlOnTkVNTQ2effZZrFmzBi+++KJR+6anp19y+pglS5YYvk5JSTGs87uQVCrFv//9b2NLJCLq5ly7Fv/9tRa5xWrkFlfj11p9yhcqd8XUwYFIVSkwLkqBSF93uDkZ/eOQiMjqGf0T7/7770dSUhK2bNkCIQS++uqryx6RS0RkTr92rOXbUFzdLeVLVnpj6uD+SFPJMVKl4Fo+IrJrRg2AOp0OCQkJOHDgAAYNGmTqmoiIjHauXYv8o/ojdjcUXZDy+bjifwZ3reWL8mPKR0TUyaifhg4ODhgyZAhOnjyJ0NBQU9dERHRFR2vPGtbyffdrLc616+AsdUBSiLdh6BsZJkeAhzMcHJjyERFdzOg/h0+dOoW4uDgMHz4c7u7uhvvXrVtnksKIiDpdmPLlFtWgpFZ/YEaIjwsy4gKRppJjbIQvov2Z8hERGcPon5TNzc345ptvDLeFEPjDH/5gkqKIiI6d7lzLd2nKd8fgfkz5iIiug9EDoEajwbhx47rdd+7cuRteEBHZp/MXpnzFNTii7p7ypYbJMS6SKR8R0Y1w1Z+ib7/9Nt566y0cO3YMCQkJhvvPnDmDtLQ0kxZHRLatM+XLLa7B1gtTPqU3Msb1Q6pKgVEqpnxERDfaVQfA++67D1OmTMEf//hHLF261HC/p6cnFAqFSYsjIttyvl2Lbce61vId7kj5lN4uuD1WP/CNi/TFAKZ8REQmddWfsN7e3vD29sYnn3zSF/UQkY05XtfScYqWGmwtqUVLuxZOUgmSlD54elzHWj6VHP2Y8hER9Rn+iU1EN1SrRottR+sMa/mKa5oBAMHeLrgtNgCpKgXGRigwwN8D7s78EUREZA786UtE162nlG9YsDd+PzYCqSo5RqnkCPR0YcpHRGQBOAASUa+1arTYfqzOMPQZUj4vF9waG4A0lQJjwhUYGMCUj4jIEvEnMxEZ5URdCzZ0HLyx9ddanG3TwpEpHxGRVeIASESX1arRYcsRtWHoK7og5UsfFIDUjqtvMOUjIrI+/KlNRAYnOtby5RbXYPMRNVradYaU76mxEUhTyTEyTI7+Xkz5iIisGQdAIjvWqtHi+2NdR+weqtanfEFezrgl3BM3DwrGmI6Uz4MpHxGRzeBPdCI7c7K+xXAi5s0l+rV8MgcJhim9MX9sONJUCowKkwNn6xAcHGzucomIyAQ4ABLZuDaNDt+X1hmGvoPVZwAA/T2dMXmgv/68fJEKDArw7JbyVbbwI14iIlvFAZDIBpXVn+u2lq+5M+UL9sb8MeFIC9enfFzLR0RknzgAEtmANo0OPxyvw4ai7ilfoKczfjPQH2k9pHxERGSf+JuAyEqVN5wznIj5wpRvaLAX5o8JN5yXr7+XK6RM+YiI6AIcAImsRLtWhx9Ku66+caBKn/L183TGxAH+SFPJMTbSFzH9mPIREdGV8bcEkQXrTPn0a/lqcaZVY0j55nWkfKlM+YiIqJc4ABJZkAtTvtziGuw/1ZXyTYj2M6R8gwI84enC/32JiOja8DcIkZlVNJ5DbpF+4NvUkfJJHSQYGuSF340OR1o4Uz4iIrqxOAAS9bF2rQ47jtcht0iNDcXVhpQvwMMJE6L9DNfYjenHlI+IiEyDv12I+kBl43nDx7qbDqvR1JHyJQZ54bejVUhTKZCqkiPImykfERGZXp8MgHl5eZg3bx60Wi1mzZqFRYsWdfv+tm3bMH/+fOzbtw85OTmYNm2a4XsnT57ErFmzUFZWBolEgg0bNkClUvVF2UTXTKPV4ccT9frz8hXX4JfKJgD6lO/mjpRvHFM+IiIyE5P/5tFqtZg7dy42bdoEpVKJlJQUZGRkIDY21rBNaGgoVqxYgWXLll2y/0MPPYRnn30WEydORHNzMxwcHExdMtE1qWw8j7zOlO+IGo3n9SnfkP5M+YiIyLKYfAAsKChAVFQUIiIiAACZmZlYu3ZttwGwM9G7eLg7dOgQNBoNJk6cCADw8PAwdblERutM+TrPy9eZ8vm7O+GmSN+Oa+z6IrafB7xcHM1cLRERUReTD4AVFRUICQkx3FYqldi5c6dR+x45cgQ+Pj648847UVpaigkTJmDp0qWQSqXdtsvOzkZ2djYAoKqqCpWVlb2qUa1W92p7W8U+6F2pD9XN7fjv8UZsLW1C/vEmnGnTwkECDPZ3xayhfhge5I7kIA/4ezrDQSIBcA7NdefQ3Hfl3zB8P+ixD13YCz32QY990LPWPph8ABRCXHKfRGLcx18ajQbbt2/H3r17ERoaiunTp2PFihV49NFHu22XlZWFrKwsAEBycjKCgoJ6Xee17GOL2Ae9zj5cmPLlFtXg546Uz8+9cy2fbad8fD/osQ9d2As99kGPfdCzxj6YfABUKpUoKysz3C4vLze6UUqlEkOHDjV8fPw///M/+Omnny4ZAIlupOrmdmwsOGlYy9dwTgOpBBgS5IUn01RIVcmRplIg2Idr+YiIyDqZfABMSUlBSUkJSktLERwcjJycHKxevdrofevr66FWq+Hv74+tW7ciOTnZxBWTvdFodfipM+UrrsHeiq6Ub2yEr+G8fHGBnjaZ8hERkf0x+QAok8mwfPlyTJo0CVqtFjNnzkRcXBwWL16M5ORkZGRkYNeuXZg6dSrq6+vx9ddf4/nnn8fBgwchlUqxbNky3HLLLRBCICkpCbNnzzZ1yWQHqprOI69YjdziGnx7pMaQ8iUEeWFmoh8mxIUgTaVAkLcLZFIeeU5ERLalT05Alp6ejvT09G73LVmyxPB1SkoKysvLL7vvxIkTsW/fPpPWR7ZPo9Vh58kGQ8q3p7wRAODr5ogx4b5IU8kxpiPlO1uvtsr1HERERMbiGWjJZlWfaTWcl+/bw2rUn2uHVALE9/fCE6lhSFUpMDpcgeCLUr6z9WYsmoiIqA9wACSbodUJ7LxgLd/uC1K+0eH6kzCPiVBgcKAXvF25lo+IiOwXB0CyatVnWrHxsP4ULRs7Uj6HjpTv8dQwpPWQ8hEREdkzDoBkVbQ6gYKTXSlfYRlTPiIiot7iAEgWr6Yj5dtQVINvj6hR13JpypcWLofS25UpHxERkRE4AJLFuVLKlxomR6pKgTERCsT3Z8pHRER0LTgAkkXoTPlyi9XYeLjGkPINDvTCY6PCkKaSY3SEgikfERHRDcABkMxCqxPYVdaA3KKOlK+8AUIACjdHjArTX2ptTIQCg/t7wYcpHxER0Q3FAZD6jLq5FRsPqzuO2K3B6ZZ2SADE9/fEnJFhhgM4mPIRERGZFgdAMhmtTqCwrOvqG7vK9Cmf3NURo1RypIbJMTbSF/FM+YiIiPoUB0C6oWo7U75i/Xn5as+2QQJgcKAnskaEGg7gUPq4wpEpHxERkVlwAKTrotMJFJZ3reUr6Ej5fFxlGBWmMFxjNyGIKR8REZGl4ABIvXb6bFvHEbs1yCvuSvniAj0xuyPlG8uUj4iIyGJxAKSr0ukEdpc3Gtby7TxZf0nKNzrcFwlBnpC7OZm7XCIiIroKDoB0WafPtuHbw2psKK7GxmI11B0pX2y/rpRvTLgCIXKmfERERNaGAyAB0Kd8v1SdxbsHjujX8p2sh86Q8umvvjE6XIEhQV5M+YiIiKwcB0A71pny6Y/YrUFNc1fK9+jw0I6rb/gilCkfERGRTeEAaEd0OoE9Ffq1fBuKulI+bxd9ypfoK8VvhqiQGOTNlI+IiMiGcQC0cXUtbdh0WI0NxTXIK9anfAAQ288Djw4P7bj6hj7lU1dXISjI38wVExERkalxALQxOp3Az5WN2NBxXr6fTnQ8gQ6qAAAgAElEQVSlfCPD5IbLrQ0J8oaCKR8REZFd4gBoA+pbutby5RWrUd3cCkCf8s0cHoLUMAXGRnItHxEREelxALRCnSlf51q+zpTPy7l7ypcYzJSPiIiILsUB0ErUt7Rh05FabCiq7pbyxQR44JGUEKSpFIa1fE4ypnxERETUMw6AFurClC+3qAY/MuUjIiKiG4QDoAVpONeOTUfU2FCkP2K36kz3lE9/jV2mfERERHR9OACakRACv1Q2Gdby/XiiHlqdgJezDCPCfPDYqDCMDldgqJIpHxEREd04HAD7WGfKl9uR8p3qSPkGBXhgRrKyYy2fAmFyN6Z8REREZBJ9MgDm5eVh3rx50Gq1mDVrFhYtWtTt+9u2bcP8+fOxb98+5OTkYNq0ad2+39TUhJiYGEydOhXLly/vi5JvmJ5SPk9nGUaG+iBrZBhGRygwNNgbvu5M+YiIiMj0TD4AarVazJ07F5s2bYJSqURKSgoyMjIQGxtr2CY0NBQrVqzAsmXLLvsYzz33HMaNG2fqUm+Yxs6Ur+PqG5VN+pRvoL87HkpSIk0lx5hIBVRyd6Z8RERE1OdMPgAWFBQgKioKERERAIDMzEysXbu22wCoUqkAAA4Olw5Du3fvRnV1NSZPnozCwkJTl3tNhBDYd6oJuR1X3/jhuD7l83CSYmSYHLNGhCJNpUBSiA9TPiIiIjI7kw+AFRUVCAkJMdxWKpXYuXOnUfvqdDo8/fTT+Pjjj7Fly5Yet8vOzkZ2djYAoKqqCpWVlb2qUa1W92p7AGhq1WL7iSZ8V9qI7443oaq5HQAQKXfGPTFyDA92xwilB4K8XDquvqFBa2MtKht7/VR95lr6YIvYBz32QY996MJe6LEPeuyDnrX2weQDoBDikvskEolR+7711ltIT0/vNkBeTlZWFrKysgAAycnJCAoK6nWdvdmnquk84v+2GZqOlG9EmByzVXKMVikwTKlfy2fsa7Q019I7W8Q+6LEPeuxDF/ZCj33QYx/0rLEPJh8AlUolysrKDLfLy8uNbtSPP/6I7du346233kJzczPa2trg4eGBpUuXmqpcowR6uWDJ5IHwd3fCmAgFVAo3OMukZq2JiIiIyFgmHwBTUlJQUlKC0tJSBAcHIycnB6tXrzZq31WrVhm+XrFiBQoLC80+/HX64y3R5i6BiIiI6JqY/BBUmUyG5cuXY9KkSYiJicE999yDuLg4LF68GOvWrQMA7Nq1C0qlEp999hnmzJmDuLg4U5dFREREZLf65DyA6enpSE9P73bfkiVLDF+npKSgvLz8io/x8MMP4+GHHzZFeURERER2hSehIyIiIrIzHACJiIiI7AwHQCIiIiI7IxGXO1GfFfPz8zNcWcRYarUa/v7+pinIirAPeuyDHvugxz50YS/02Ac99kHP0vpw/Phx1NbWXnU7mxsAr0VycrLFXmauL7EPeuyDHvugxz50YS/02Ac99kHPWvvAj4CJiIiI7AwHQCIiIiI7I33hhRdeMHcRliApKcncJVgE9kGPfdBjH/TYhy7shR77oMc+6FljH7gGkIiIiMjO8CNgIiIiIjvDAZCIiIjIzljcAFhWVobx48cjJiYGcXFxeOONNwAAdXV1mDhxIqKjozFx4kTU19cDAIqLizFq1Cg4Oztj2bJl3R6roaEB06ZNw6BBgxATE4Mff/zxss+Zl5eHgQMHIioqCkuXLjXcv3XrVgwbNgyDBw/GjBkzoNFoLrt/aWkpRowYgejoaEyfPh1tbW0AgBUrVsDf3x+JiYlITEzEu+++a9N9WL58OaKioiCRSLqdg6i+vh5Tp05FQkIChg8fjgMHDhjdB3P1YubMmQgICMDgwYO73d/Tc16sp/fEv/71L8THxyMxMRGjR4/GoUOHbLoPPb0n/vKXvxj+vxg8eDCkUinq6upstg+PPvoohgwZgoSEBEybNg3Nzc0AgG3btmHYsGGQyWRYs2aNUa/fEvvw2WefIS4uDg4ODlc8HcbVtjt58iQ8PDwuqc8We7FgwQIMGjQICQkJmDp1KhoaGgAA7e3tmDFjBuLj4xETE4NXXnnFKvvQ0+u7WE/92rRpE5KSkhAfH4+kpCRs3brVpvtwpe327duHUaNGIS4uDvHx8Th//rzRvbgiYWEqKyvF7t27hRBCNDU1iejoaHHw4EGxYMEC8corrwghhHjllVfEwoULhRBCVFdXi4KCAvGnP/1J/OUvf+n2WA899JB45513hBBCtLa2ivr6+kueT6PRiIiICHH06FHR2toqEhISxMGDB4VWqxVKpVIcPnxYCCHEc889J959993L1nz33XeLTz75RAghxJw5c8Rbb70lhBDigw8+EHPnzrWbPuzZs0eUlpaKsLAwoVarDfc/88wz4oUXXhBCCFFUVCRuvvlmi+6FEELk5+eL3bt3i7i4uG739/ScF+vpPdHY2GjYZu3atWLSpEk23Yee3hMXWrdunRg/frwxLRBCWGcfLvx3f+qppwz7lJaWil9++UU8+OCD4rPPPjO6B5bWh0OHDoni4mIxbtw4sWvXrh5rvtp2d955p5g2bdol9V2NNfZi48aNor29XQghxMKFCw21rVq1SkyfPl0IIcTZs2dFWFiYKC0ttbo+9PT6LtZTv/bs2SMqKiqEEELs379fBAUFGdUDa+1DT9u1t7eL+Ph48fPPPwshhKitrRUajcboXlyJxQ2AF8vIyBDffvutGDBggKisrBRC6P9xBwwY0G27559/vts/XGNjo1CpVEKn013x8Xfs2CF+85vfGG6//PLL4uWXXxY1NTUiMjLScP+2bdvElClTLtlfp9MJX19fwz/chY93PQPgxSy9Dxe6+Jd9enq62L59u+F2RESEqKqquuJjXImpe9GptLT0kv+Zr/acQlz5PXGh1atXi8mTJxtVy+VYeh8udKUB8N577xXZ2dlG1XI51tQHnU4nHnvsMbF06dJu98+YMaPXA+DFzNmHTlcbeq603ZdffimeeeaZS+q7FtbUCyGE+OKLL8R9990nhND/XLjttttEe3u7qK2tFdHR0eL06dNGPc7FLKEPQnR/fT25Ur90Op1QKBTi/PnzRtVzMWvqw8XbrV+/Xtx///1GPX9vWdxHwBc6fvw49u7dixEjRqC6uhr9+/cHAPTv3x81NTVX3PfYsWPw9/fHI488gqFDh2LWrFk4e/bsJdtVVFQgJCTEcFupVKKiogJ+fn5ob283RNJr1qxBWVnZJfufPn0aPj4+kMlk3fbv9Pnnnxs+9rnc/sawhj5cyZAhQ/DFF18AAAoKCnDixAmUl5f36jE69UUvrsSY57zae+Kf//wnIiMjsXDhQrz55pu9ev5O1tAHY7S0tCAvLw933XXXNe1vTX145JFHEBgYiOLiYvz2t7/t1fNcjbn7cL3Onj2LV199Fc8///x1P5Y19uL999/HlClTAADTpk2Du7s7+vfvj9DQUDzzzDNQKBS9fkxL6sOFr+9afP755xg6dCicnZ17va819uHC7Y4cOQKJRIJJkyZh2LBheO211675+S9msQNgc3Mz7rrrLvz973+Hl5dXr/fXaDTYs2cPHn/8cezduxfu7u7d1rV1Epc5C45EIoFEIkFOTg6eeuopDB8+HJ6enoZf6MbsDwC33347jh8/jn379mHChAmYMWNGr1+HtfThShYtWoT6+nokJibiH//4B4YOHdrrxwD6rhfX60rvCQCYO3cujh49ildffRUvvvhirx/fWvpgjK+//hppaWnX9AvO2vrwwQcfoLKyEjExMfj0009v2ONaWx8u5/nnn8dTTz0FDw+P63oca+zFSy+9BJlMhvvvvx+A/o9kqVSKyspKlJaW4q9//SuOHTvWq8e0pD5c/Pp66+DBg/jDH/6Af//7373e1xr7cPF2Go0G33//PVatWoXvv/8eX375JbZs2XJNNVzMIgfA9vZ23HXXXbj//vtx5513AgD69euHU6dOAQBOnTqFgICAKz6GUqmEUqnEiBEjAOj/qtqzZw/KysoMi8//9a9/QalUdku0ysvLERQUBAAYNWoUtm/fjoKCAowdOxbR0dEAgEmTJiExMRGzZs2Cn58fGhoaDAdGXLi/r6+v4S+W2bNnY/fu3Tbbhyvx8vLCBx98gJ9//hkfffQR1Go1wsPDLbYXV9LTcxr7nrhQZmYmvvrqq150wbr6YIycnBzce++9Rm17IWvtg1QqxfTp0/H555/3+jVfjqX0oSePPPIIEhMTkZ6efsXtdu7ciYULF0KlUuHvf/87Xn75ZSxfvrxXz2WNvfjwww/xzTffYNWqVYY/ElevXo3JkyfD0dERAQEBSEtL69V1Zi2pD5d7fca+JwD9z86pU6fio48+QmRk5FW3v5A19uFy2ymVSowbNw5+fn5wc3NDeno69uzZ06te9KT3MYyJCSHw6KOPIiYmBr///e8N92dkZODDDz/EokWL8OGHH+KOO+644uMEBgYiJCQEhw8fxsCBA7FlyxbExsYiJCQEP//8s2E7jUaDkpISlJaWIjg4GDk5OVi9ejUAoKamBgEBAWhtbcWrr76KZ599FgCwcePGbs81fvx4rFmzBpmZmd1qO3XqlCFuXrduHWJiYmy6Dz1paGiAm5sbnJyc8O6772Ls2LG9+musr3txJT09p7HviZKSEsMAvX79esPXttqHK2lsbER+fj5Wrlxp9D6A9fVBCIGjR48iKioKQgh8/fXXGDRoUK9e8+VYUh968sEHHxi13fbt2w1fv/DCC/Dw8MCTTz5p9PNYYy/y8vLw6quvIj8/H25ubob7Q0NDsXXrVjzwwANoaWnBTz/9hPnz5xv1HJbUh55en7HviYaGBtx666145ZVXkJaWZtQ+nayxDz1tN2nSJLz22mtoaWmBk5MT8vPz8dRTTxn13FdlkpWF12H79u0CgIiPjxdDhgwRQ4YMEevXrxe1tbXi5ptvFlFRUeLmm282LIo9deqUCA4OFp6ensLb21sEBwcbjrjbu3evSEpKEvHx8eKOO+4QdXV1l33O9evXi+joaBERESFefPFFw/3PPPOMGDRokBgwYID429/+1mPNR48eFSkpKSIyMlJMmzbNsFB10aJFIjY2ViQkJIibbrpJFBUV2XQf3njjDREcHCykUqno37+/ePTRR4UQ+oMgoqKixMCBA8XUqVN7fH5L6kVmZqYIDAwUMplMBAcHG4587uk5L9bTe+J3v/udiI2NFUOGDBE33XSTOHDggE33oaf3hBD6g6Q6j3bsDWvrg1arFampqWLw4MEiLi5O3HfffYbnLygoEMHBwcLNzU0oFAoRGxtrlX344osvRHBwsHBychIBAQGXPejJ2O2u5SAQa+xFZGSkUCqVhnrnzJkjhBDizJkzYtq0aSI2NlbExMSI1157zSr70NPru1hP/frzn/8s3NzcDPsPGTJEVFdX22wfrrTdxx9/LGJjY0VcXJxYsGCBUT0wBi8FR0RERGRnLHINIBERERGZDgdAIiIiIjvDAZCIiIjIznAAJCIiIrIzHACJiIiI7AwHQCKiG+SFF17AsmXLzF0GEdFVcQAkIiIisjMcAImIrsNLL72EgQMHYsKECTh8+DAA4M0330RsbCwSEhKQmZlp5gqJiC5lcZeCIyKyFrt370ZOTg727t0LjUaDYcOGISkpCUuXLkVpaSmcnZ3R0NBg7jKJiC7BBJCI6Bpt374dU6dOhZubG7y8vJCRkQEASEhIwP3334+VK1dCJuPf2URkeTgAEhFdB4lEcsl969evx9y5c7F7924kJSVBo9GYoTIiop5xACQiukZjx47Fl19+iXPnzuHMmTP4+uuvodPpUFZWhvHjx+O1115DQ0MDmpubzV0qEVE3/GyCiOgaDRs2DNOnT0diYiLCwsIwZswYSCQSPPDAA2hsbIQQAk899RR8fHzMXSoRUTcSIYQwdxFERERE1Hf4ETARERGRneEASERERGRnOAASERER2RkOgERERER2hgMgERERkZ3hAEhERERkZzgAEhEREdkZDoBEREREdoYDIBEREZGd4QBIREREZGc4ABIRERHZGQ6ARERERHaGAyARERGRneEASERERGRnbG4AnDx5srlLICIiIjILY+cgiRBCmLiWPuXt7Y3o6Og+ea729nY4Ojr2yXPZG/bWtNhf02FvTYv9NR321rT6qr8lJSVobGy86nYyk1fSx6Kjo1FYWNgnz1VZWYmgoKA+eS57w96aFvtrOuytabG/psPemlZf9Tc5Odmo7WzuI2AiIiIiujIOgERERER2hgMgERERkZ3hAEhERERkZ2zuIBAiIlNo1+pQ19KO02fbcLqlreO/F95ux+mWNng4aPDOff3g4ig1d8lERD3iAEhEdkUIgeZW7QVDXNfwdulQ14baljbUnW1HU6umx8d0lErg7eIILxcZjp1uwTmxB2seTunDV0VE1DscAInIamm0OtSfa0ft2Z4TOcNgd7YNtS3tqG9pQ5u259OfejhJ4e3qCG8XGXxcHBET4AFvF/3tzvu9XRwhd3VEgKcz+nk6QeHqBBdHKVxkDnj689349+4q/PW/R/H0TZF92A0iIuNxACQisxNCoKVNe/k07uKkriOVO322DY3ne07lZA4Sw9Dm4yJDgIczov09DAOct6v+vz4uMvh5OCHAwxkBHk7wcHaEi8wBLo4OcJI6QCKR9Oq1PDtWiaJ6Df64oQjDQ30wJsL3ettDRHTDcQAkohtKqxOob7lKGtf5vc6PWFva0arR9fiY7k7SrsHNRYZoP3ckK70Ng5yPi6N+mHPVD3oBns7wc3eEq6MMzjIHuMgcIJP2zTFvUgcJvng4BQnL/ou7P9qNfU+PRYCnS588NxGRsTgAElGPWto0V0/kOtfKdXy/8Vw7evqAVdqZynUMcwo3J4T7uhkGO58LPmL1dXdCgIcT+nk6w8vF0TDIOct6n8r1NV93J6ybORxp//gBt7+3Cz/8Nq3PBlAiImNwACSyMyfqWpB/rBGo1F5+3dwFH7Gev0Iq5+Yo7bYuLtLXHcOUnWlc1/0+ro7w93BGoIcz/Dwc4SKTwsVRCmeZAxxteChKCvHBP6YORtaafXjyy/3417Qh5i6JiMiAAyCRHTmibsaQZfndBjupBPC6cGhzlSFU7gofV9klBz/oUzlnBHg4w9tFZhjknKUOcHCw7FTOHGaPCsOO43X4948nkaZS4MHkEHOXREQEgAMgkd0QQmDu5/shc5Dg9YlKDAoLRj9PJ/i5O8HVUdqRzNl2KmcO/7o7AXsrmvDYmv1IDPZGfH8vc5dERMQrgRDZi89+OYXNJbV4Ik2FuwYHYEpMAIYpfRAqd4O/hzM8XWQc/kzAWSbFN7OGw1nmgIz3CnDmCkcuExH1Ff60J7IDZ85r8NTaAxgU4IHZI8Ig48e1fUrp44r/PJSEkw3ncOeKXRCi5/MQEhH1BbMOgHl5eRg4cCCioqKwdOnSS77/+uuvIzY2FgkJCbjllltw4sQJM1RJZP1e+PYwTjW1YtH4SET5u5u7HLs0YYA//jx5IDaX1OK5vGJzl0NEds5sA6BWq8XcuXORm5uLQ4cO4ZNPPsGhQ4e6bTN06FAUFhZi3759mDZtGhYuXGimaoms177KJryxvRRT4wMxNaG/ucuxa3+8JRq3xfbDK1t+xfpD1eYuh4jsmNkGwIKCAkRFRSEiIgJOTk7IzMzE2rVru20zfvx4uLm5AQBGjhyJ8vJyc5RKZLV0OoHHP98HT2cp5qap4OXiaO6S7JpEIsHq+4dBJXfDg6v34nhdi7lLIiI7ZbYBsKKiAiEhXadEUCqVqKio6HH79957D1OmTOmL0ohsxoeFZdhxvB7zRkdgXKSfucshAJ4uMmyYPQJtWh1ue7cArRqtuUsiIjtkttPAXG4RdE9n91+5ciUKCwuRn59/2e9nZ2cjOzsbAFBVVYXKysobV+gVqNXqPnkee8TeXr+6cxo8vfYABvu74naVI6qrThm+x/6ajjG99QTw+qQwzPn6GKa/vwNv3RZp+sJsBN+7psPempal9ddsA6BSqURZWZnhdnl5OYKCgi7ZbvPmzXjppZeQn58PZ2fnyz5WVlYWsrKyAADJycmXfRxT6cvnsjfs7fX5v89+QVOrFm9Pi0HSoOBLvs/+mo4xvc0KCkJRo8Dft5Xi5pjz+N2YiD6ozDbwvWs67K1pWVJ/zfYRcEpKCkpKSlBaWoq2tjbk5OQgIyOj2zZ79+7FnDlzsG7dOgQEBJipUiLrs/NEPd7ZeRKZQ4NxW2w/c5dDPfjLbbFIVcmx4OtD2FFaZ+5yiMiOmG0AlMlkWL58OSZNmoSYmBjcc889iIuLw+LFi7Fu3ToAwIIFC9Dc3Iy7774biYmJlwyIRHQpbceBH35uTpibpoK7My/4Y6lkUgd89UgK5G5OmPZhIWqbW81dEhHZCbP+ZkhPT0d6enq3+5YsWWL4evPmzX1dEpHVe3vHceytaMLS9EEYESo3dzl0Ff4ezlj3SApG//MH3PZeAX747WhIeaJuIjIxXgmEyIZUNZ3Hs7nFGBnqgweTlXDgIGEVhofJ8cYdg7HzZAN+9+UBc5dDRHaAAyCRDXnm60M4167FwvFRCPJ2NXc51AuPp6nwwLBgvLXjOFbv4TlPici0OAAS2Yjvfq3Fqj0VeDg5BJNjeNCUNXrnniEYHOiJrM/24WBVk7nLISIbxgGQyAa0aXR44vP9CPZywWOjwuDqKDV3SXQNXBylWD9rOBylEmS8vwvNrRpzl0RENooDIJEN+Nu2YyiuacaC8ZEYqvQ2dzl0HULlbvj0wWQcr2vBtA8LL3vSfCKi68UBkMjKnahrwZJvj2B8pC/uGxbc4xV1yHr8ZqA/nv/NAGw8rMb/fXvE3OUQkQ3iAEhk5eavPQidEHh6XAT8PS5/tRyyPs9NHIDJg/zx4qYjyCuqNnc5RGRjOAASWbFvDlXjqwNVmD0yFBMG+pu7HLqBJBIJPn0wCSE+rrh/9V6crG8xd0lEZEM4ABJZqZY2DX77xX5EKNyQNSoMzjIe+GFrvFwcsX7WCJxr1+K2dwvQptGZuyQishEcAIms1CtbfsXx+nNYOD4SgwO9zF0OmUhsoCfeuycR+6vO4JFPfzZ3OURkIzgAElmhI+pmvPbdUUwZFIB7EoPMXQ6Z2L3DgvFkmgqr91Tgn9+XmrscIrIBHACJrIwQAnM/3w8nqQTzxoRD7uZk7pKoD/ztjjiMDPPB0+sOoeBkvbnLISIrxwGQyMp89sspbC6pxRNpKtwc7WfucqiPyKQOWPvIcPi4ynDnB4U4fbbV3CURkRXjAEhkRZrOt2P+2gMYFOCB2SPC4Cjl/8L2JMDTGV8+koKas63IeH8XdDqeJJqIrg1/exBZkRc2HkFVUysW3RyFKH93c5dDZjBKpcBfM+Kw43g9nlp70NzlEJGV4gBIZCV+qWzEm9+X4s74QEyNDzR3OWRGT6apkJkYhDe/L8WneyvMXQ4RWSEOgERWQKcTeOLz/fB0luKJNBW8XBzNXRKZkUQiwfuZiYjt54FZ//kFxdVnzF0SEVkZsw6AeXl5GDhwIKKiorB06dJLvr9t2zYMGzYMMpkMa9asMUOFRJbhw8Iy7Dhej3mjIzAukgd+EODqKMX6WSMgdZDg9vcLcLZVY+6SiMiKmG0A1Gq1mDt3LnJzc3Ho0CF88sknOHToULdtQkNDsWLFCtx3331mqpLI/E6fbcOCrw8hMcgLjwwPgdRBYu6SyEKoFG745IFhOFrbgns+3g0heFAIERnHbANgQUEBoqKiEBERAScnJ2RmZmLt2rXdtlGpVEhISICDAz+pJvv1pw1FaDjXjkU3RyFM4WbucsjCTInph/+dEI0NRTV4eXOJucshIishM9cTV1RUICQkxHBbqVRi586d1/RY2dnZyM7OBgBUVVWhsrLyhtR4NWq1uk+exx6xt3q7K5vxzk8nMS1GjmSF7oa9t9lf0zFHb7MSPLGtxBPPbzyMaA8tRod593kNfYXvXdNhb03L0vprtgHwch9VSCTX9tFWVlYWsrKyAADJyckICuq7S2P15XPZG3vvrUarw/M52+Hv4YSnJsQiMkxxQx/f3vtrSubo7drZAUhYlo8nNpzAL0+PQ7CPa5/X0Ff43jUd9ta0LKm/ZvtsValUoqyszHC7vLzcohpDZG5v7ziBvRVN+P3YCIwIlZu7HLJw3q6O+ObRFJxt0+K29wrQrtWZuyQismBmGwBTUlJQUlKC0tJStLW1IScnBxkZGeYqh8iiVDWdx//mFWNkqA8eSlbCgQd+kBHig7zxzt1D8HNlE2Z9+ou5yyEiC2a2AVAmk2H58uWYNGkSYmJicM899yAuLg6LFy/GunXrAAC7du2CUqnEZ599hjlz5iAuLs5c5RL1qWe+PoRz7VosHB+F/t62+1Ee3XgPJCvx2KgwfLS7HP/+8bi5yyEiC2W2NYAAkJ6ejvT09G73LVmyxPB1SkoKysvL+7osIrP67tdarNpTgdkjQjE5JsDc5ZAVenPqYOwub8S8rw4iSemN5BAuISCi7nh+FSIL0qbR4YnP9yPY2wWPjQqDq6PU3CWRFXKUOmDdzBR4Ocsw9YNC1Le0mbskIrsmhIBGZ1nn6eQASGRBXs8/iuKaZiy8KRJDlbZ7Kg8yvUAvF3zxcDKqzrTijvd3QWdhv3yI7IEQAluOqDF6+Q/4Z8Epc5fTDQdAIgtxoq4FSzYdwfhIX9w7LPiaT4tE1Gl0hC9euy0G20vrsOCbQ1ffgYhumO3HTmP82z9iwr9/wtHas5BY2JV6zLoGkIi6zPvqAIQAnh4XAX8PZ3OXQzZi/tgI7Dhej7/lH0NqmBx3DeHptohMaeeJejyXV4xNR2rh6+aIZ26KwAPDguGHs+YurRsOgNdoX2UTfPiRCt0g3xyqxtqD1fjtaBUmDPQ3dzlkQyQSCT68NxH7TzVh5qe/IL6/FwYEeJvwJucAACAASURBVJi7LCKbs6e8AYvzDmN9UQ18XGWYPyYcDyQpkRjsDamDBJWVLeYusRt+BHwNjte1YMQb2/HMxuO8+Dpdt5Y2DX77xX5EKNwwZ1QYnGU88INuLDcnGTbMHgEAuP39ArS0acxcEZHt2H+qCXeu2IWkv23H9mN1eDJNhbzZI7EsIw5JIT6QWuh5XDkAXgOVwg1z01T47FAd/rSh2NzlkJV7ZcuvOF5/Dn+4ORJxgV7mLodsVISvO1bdPxQl6rO4d+Ue/vFKdJ2Kq88g8+PdGPLXfHx7WI05I0OxYfZw/O2OOIwIk1vs4NeJHwFfo9dui8WhitNYuvVXqDqSG6LeOlzTjNe+O4r0mADczbVZZGK3xQXij7dE4eUtv+LVrb9i0S3R5i6JyOr8WnsWS749glV7yuEik+KRlBA8lKREargCjlLrydU4AF4jBwcJ/nlrBDK/LMWTX+xHqI8LpsT0M3dZZEWEEHjyi/1wkkowb3Q45G5O5i6J7MCfJw/CTyfq8VzeYYwI9cH4aK45JTLGiboW/HlTCVYUlkHmIMH9w5SYkazEmAhfOMmsZ/DrZH0VWxBnmQM2Zo1EqNwVd3+0Gz9XNJq7JLIi//m5EptLavFEmgrjo/3MXQ7ZCQcHCdbMSEY/T2dM/3gPKhvOmbskIotW0XgOT3y+D9FLt+Lj3WW4J6E/1j6Sguy7E3DLAH+rHP4ADoDXzcfVEd89PgpujlJMzv4J5Q2WdZQPWaam8+14at1BDArwwOwRYVb1sQFZP7mbE9Y/OhxNrRrc/n4B2rU6c5dEZHGqms5j/lcHEPnyVrzz00ncEReILx9OwXuZiZg0KAAuVn6lJv7WuQFC5W7Imz0CZ1q1uPntn9B0vt3cJZGFe2HjEVQ1tWLRzVGI8nc3dzlkh4YEe+Nf0+Kxp6IJj63ZZ+5yiCyGurkVC74+hIiXt2D596WYPNAfX8xIxof3JiI9tp/NXKKTawBvkGEhPvjPQ0m44/0CTMr+CdvmpjHVocv6pbIRb35fijvjAzE1PtDc5ZAdezglFDtK6/HOzpNIVcnx6AgezEb2q66lDX/971G8ub0ULe1aTB4UgNkjQjFhgD88nG1vXLK9V2RGt8b2w/I74/H45/uR+fFurJmRzMt5UTc6ncATn++Hl7MMc9NU8HJxNHdJZOeW3xmP3RWN+O2XB5AU7INEXoOa7EzjuXb8fdsxvL7tGJrOazBxgB/mjAzDbwYGwNPFdsckRlQ32GOpKiwcH4kv9lfhmXW89iZ1t2JXGXYcr8e8MeEYG8kDP8j8nGQO+ObR4XBzlOKOD3ahoaXN3CUR9YnmVg1e2VKC8Je24IVvjyAp2Bs5DwzDZw8l464hQTY9/AFGDoDLly9HfX29qWuxGUtvjcH0xCC8vu0Y/rH9mLnLof9v777jqqz7x4+/GCIbQYaHIYjiYi8RcSuSWai5cu+VtrOsu19aafHNhq3bkdvKTHOlmTPTxIWKO0UFTQFlCSqojM/vD+vcejtuRA4HOO/n4+Hj4TnnOtfnfd5cHN7XdX1GJZF1/Ravrz1OkKstQ8M9Kv0kocJwaGzNWTEknIt5N+i+IIESWeZSVGP5t4r4ZNsZvKdu4a1f/sS3jg3f9gvmpyFh9Al2w87CMO7MlKoATE9PJzw8nN69e/Prr7/KDPL/g5GREYv7BdOyngMvrz7G6qNp+g5JVAJv/XKCKwWFTGzfgLoOlvoOR4i7tK5fm7gnG7PtTBZvrjuh73CEKHc3i4r5ckcy9T/Yyms/H8e7tiXz+wSyamg4/UPdDW4u1lIVgFOmTCEpKYnhw4ezYMECfHx8eOuttzhz5sxjNf7rr7/SqFEjGjRoQFxc3D2v37x5kz59+tCgQQMiIiJISUl5rPYqUg0TY9aNaEZ9Ryv6fXuQveflCqoh233udkf7vsFuPNVUJgwXldOrbevT3b8O07adYdUROXEV1cOtohJm7UqhwQdbeWHVUVxtzfmmVwA/D2/GkGZ1qW1lWIXfP0rdB9DIyIg6depQp04dTE1NycnJoWfPnrz++utlari4uJhx48axfv16jh8/zpIlSzh+/O4+c3PnzsXe3p7Tp0/z8ssv88Ybb5SpLX2xNa/B1rGR2NQ0ocucvaRkyRyBhqiouITnfjqMk5UZz0V5YVUNR5OJ6sHIyIjFfYNp4GjFkB8SOZ1xTd8hCVFmRcUlzN97nkZxWxmz/AgOljX49zN+rBvRjBHNPXGyrqnvEPWqVAXgF198QWhoKK+//jpRUVEcOXKEGTNmsH//fn766acyNbx3714aNGiAt7c3ZmZmPPvss6xevfqubVavXs3gwYMB6NmzJ1u2bKlyt5/d7CzYNLo5NwqL6TBrF1cKZI5AQzMj/hwHL+bxShtvIura6zscIR7KqqYpv4yMoLhE8fS8fRQUFus7JCEeSXGJ4rv9F2j60TaGLT2EpZkJn3f1Zf3ICMZG1aOOrbm+Q6wUSnUpIjMzkxUrVuDpefccUcbGxqxdu7ZMDV+8eBEPDw/tY3d3d/bs2fPAbUxNTbGzsyMrKwtHx7tHT86ePZvZs2cDt/srpqamlimmR5WRkVGq7WoDc2K9GbAiibZf/s7KZxtTs4ouHVNRSpvbyu7StUL+9ctxwjSWdHY3JT29ctxWqy75rYyqQ24tgS86ezFs9Rl6ztnJ7Nj6lWZKq+qQ38qqque2RCnWnbrCJ/GpJGXfwLuWGe+2ceXphva42JrD9RxSr+uvO1Zly+9DC8Ds7GwAXnrppbse/8PBwYEmTZqUqeH7Xcn77y+Y0mwDMGrUKEaNGgVAWFgYrq6uZYqpLErbVh9XV/JNLBm29BAvbEpl9dBwjGUU6ENV5M9RVyZ8e4AbxYq3OjUloKFG3+HcpTrkt7KqDrkd6urKiVyYtu0MS07d4LV29fUdklZ1yG9lVRVzq5RizbFLvPPrSQ6n5VHPwYK4Lo3pE+SKl0PlWmmpMuX3oQVgaGgoRkZGDyzEzp4t+xQn7u7u/PXXX9rHFy5cuCcx/2zj7u5OUVERubm5ODg4lLlNfRvarC7J2fm8vymJF1Ye5ase/voOSejQ1qRMvj94kZERdXmiibO+wxHikX3YpQl7z1/hrfUnaOZZi9betfUdkhBaSil+/fMy72w4ScJfudStZcGUJxrxbLAb9R0rV+FXGT20AExOTn7ga4/bFy88PJykpCSSk5Nxc3Pjhx9+4Pvvv79rm9jYWBYuXEhkZCTLly+nffv2leY2RFm9G9OI5Kx8vo5Pwau2Ja+1rTxn1aL83CoqYdyKI7jZmTMm0rParB0pDIuJsRErhoYRMO13ei9MIPHVNtJ/SuidUootSZm88+tJdp3Lwc3WnEnRDekb4kpDJ+sqXydUlFJ1RHvnnXfuelxSUsKAAQMeq2FTU1O++uorYmJiaNKkCb1798bX15d33nmHNWvWADB8+HCysrJo0KABn3766X2niqlqjIyMmP9sEO0a1OaNtcdZdqhi+iuKivXp72f48/I1Xm9bn2BZWktUYQ6WZqwZHs6VgiKenruXouISfYckDNj2M1m0/Xc80bN2czbrOm91aMCG0RFMimlII2cbKf4eQakGgZw/f54PP/yQN998k5s3b9KrVy9CQkIeu/Enn3ySJ5988q7n3nvvPe3/zc3NWbZs2WO3U9mYmhizZlgzIj7fwaDvD+Jma06LelX31ra427nsfN7bdIp2DWrTN8RNvpBElRfiXouve/gx4sfDPLfiCLN7Beo7JGFgdp/L4f+t/5PNSZk4WpkxoW19Boa641vHRvrTl1GprgDOnz+fI0eO8OGHH/L000/Trl07Jk+erOPQqjfrmqZsGROJg2UNnp63lySZb6vaeHHVUQBebVPf4OeZEtXH8AhPhoa7883u8yzYe17f4QgDsf+vK3SZs4fIL/5g/4VcXmpdjw0jI4jr0gR/V1sp/h7DQwvAAwcOcODAAQ4ePMiLL77I0qVL8fHxoU2bNhw4cKCiYqy26tias3lMJEXFio6zdpN1XRZhr+p+PpbO6mOXGBFRl44NHf/3G4SoQmb2DCTI1ZZxK45y6GKuvsMR1djh1Dy6z99H2PQd7EzOZnyUF7+OjODjp30J8aglhV85eOgt4FdfffWux/b29hw/fpxXX30VIyMjtm7dqtPgDEETFxtWDwsnZvZuOszcxa4XWsqAgSoq/1YRL6w8ireDJaMjPalpKj9HUb2YmRqzdkQzAj7+na7z93Ho1dbYWRjmMlpCN05cusrkDaf48VAqVmYmjG5el4Fh7kTUtcfURObPLU8PLQB/++23iorDoLVt4MjcPkEM/P4g3ebvY/2ICDm7qYI+3HKalJwCZvX0x7eOrb7DEUIn3OwsWDYojE6zdvHMgv1sHtNc+rmKx3Y68zrvbjzJ9wcuYm5qwrBmHgwKdadFPQdqSOGnE6XK6qVLlxg+fDidO3cG4Pjx48ydO1engRmaAaHuTOnciI0nMxi9/LC+wxGP6OTla3z02xmebOJMr8DKM9GnELrQ3seRKZ0bs/V0Jm+v/1Pf4YgqLCU7n+FLE2n8f7+x7FAa/UPcWTMsnBk9AmjTwFGKPx0qVWaHDBlCTEyMdom1hg0bMn36dJ0GZoje6uDD8GZ1mbPnPB9sPqXvcEQpKaUYv+IIZiZGvNiyHvaWcktMVH9vtG9ArK8LcVtPs/ZYur7DEVXMhSsFjF1+mIZxW/l2/wV6B2hYPTScb3oH0KGhE2ayXKrOlSrDmZmZ9O7dG2Pj25ubmppiYiL9m8qbkZERs3oF0KmhE2+vP8l3+y/oOyRRCj8mprI5KZNxUV6085GBH8IwGBkZ8V3/EOo5WDJwyUGSs/L1HZKoAtLzbvDiqqM0+HArc/acp6tvHVYOCWfus0HENHaWvtMVqFQFoJWVFVlZWdp+Hrt378bOTia31QUTYyNWDg0jQGPL8KWH2HY6U98hiYfIu1HIy2uO0cTZmhERnnK7QhgU65qm/DIygsJixVNz93CjsFjfIYlKKuPaTSb8fBzvD7bw9R/JPNHIiZVDwljYN4gnm7rI4Ec9KNVfq08//ZTY2FjOnDlDVFQUgwYN4ssvv9R1bAbL0syUTWOa42xjRrf5+zhx6aq+QxIPMGnDSdLzbjKxfQMaOMnak8LwNHSyZmHfII5fusag7w/qOxxRyWTn3+KtX05Qb+oWPv39DO0bOLJ8cBjf9g/hKd86WJqVaj0KoQOlynxISAi///47J0+eRClFo0aNqFGjhq5jM2hO1jXZMiaSZp/vIHrWbva/1AoXWYOzUjmUmssXO5J5xr8O3fzr6DscIfSmR4Arr7TO4dPtZ5n++xleaiNrnBu63IJCPtt+ls+2n+XqjSKiGzoxqnldOjVyxsZcir7KoFRXAPPz84mLi2P69On4+fmRkpLC2rVrdR2bwfNxsmbt8Agyr9+iw8zd5N8q0ndI4m8lJYrnfjqCnXkNxrf0wtZcToiEYfvo6aa0rOfAG+tOsDM5S9/hCD25drOIDzYnUW/qFt7deIpQNzuWDAjhx0Gh9Ah0leKvEilVATh06FDMzMzYtWsXAO7u7rz99ts6DUzcFlXPgcX9gjh+6SpPzd1LcYnSd0gCWLDvL+JTcnixVT1aecvADyFMjI1YOSQMB0szei7cT8a1m/oOSVSg/FtFfPzbGepN3cK/1v+JXx0bvu0XzE9DwugT7IadhZwkVzalKgDPnDnD66+/rr3ta2FhgVJSiFSUXoFufPRUU347ncXQHxIl93qWdf0Wr689TrCbLcOaeWAik3YLAYCjdU1+Hh5Odv4tOWE1EDcKi/lix1nqf7CVCWuP06C2JQv6BLJyaDj9Q91lWqxKrFTXYs3MzCgoKNCOAj5z5gw1a8oi9xXptXb1Sc6+zr/jz+HlYMF7TzTWd0gG681fTnCloJCvn/HHw95S3+EIUamEedjzRTc/xvx0hOdXHuHfPQL0HZLQgVtFJczbe56pm5O4kHuDUHc73n+iEV396uBkLfVBVVCqAvDdd9/liSee4K+//qJ///7s3LmTBQsW6Dg08d++7O7P+SsFvL8pCS97S4ZF1NV3SAZn97kcvtl9ngEhbjzV1EXf4QhRKY1u4cXOlBxmxJ8jysuB/qHu+g5JlJOi4hIWJVzg/U2nSMkpIEBjy786+tDdX4OLjRR+VUmpCsBFixbRpUsXevbsibe3N59//jmOjtLvqaIZGxuxbFAYLb/ayZifDlO3lgUdGznpOyyDUVRcwtjlh3G2NuO5KC+sakpnZiEe5JveARxKzWX08sMEutrip5H1sauy4hLFkoMXeXfjSU5n5tPUxZovuvnSI0CDq52FvsMTZVDqQSA3btxgzZo1vPDCC4wePZrPP/9c17GJ+zCvYcLG0c1xszXnmYUJHEnN1XdIBmNG/DkSU/N4pY03EXXt9R2OEJVaTVMT1g6PwMzEmK7z9nHtpsxiUBWVlCh+TEzFf9o2Bn5/EGOM+OTppmwYFcHzrbyl+KvCSlUAtm/fnn/961+8//77jBgxgoSEBGbMmFHmRrOzs4mOjsbHx4fo6GhycnLuu90TTzxBrVq1eOqpp8rcVnXkYGnG1rEtMDMxImb2HlJzC/QdUrWXlneDt3/9k0hPewaFumMsAz+E+J887C1YOiiUlJx8nlmwTwawVSFKKVYdSSP40+30Wbyfm0UlxHVpzIbREbzStj7utaT/c1VXqgKwQ4cOREVFsXTpUho1asS+ffv4888/y9xoXFwcHTp0ICkpiQ4dOhAXF3ff7SZMmMDixYvL3E51Vq+2Jb+MiCCnoJAOM3fL2bWOvbbmODcKi3m9XX00csYrRKlFN3RickwjNp3KZNKGk/oOR/wPSil+OXGJ8Ok76L4ggSsFhUx5ohEbRzfnjfY+eDnIikfVRakKwICAAMzMzDh69CiHDx/m6NGjFBSU/arT6tWrGTx4MACDBw9m1apV992uQ4cO2NjYlLmd6q6Zpz1LBoRwKuManb/ZQ1Fxib5Dqpa2JmXy/cGLDA7zIKaxs77DEaLKebujD082ceaDzUmsP3FJ3+GI+1BKsf1cHi2+3EmXOXtJz7vJpOiGbBoTwVsdfajvKIVfdVOqXuyfffYZANeuXWP+/PkMHTqU9PR0bt4s20Sfly5dQqPRAKDRaLh8+XKZ9vOP2bNnM3v2bADS09NJTU19rP2VVkZGRoW08zDNasN77dx5e+sFes7dydddvLXT9VRllSG3ALeKSxi19Dga6xo862NBTsYl7t9hoWqpLPmtjiS39/dZR1eiU6/Qb/F+Ng5qgptt2UaMSn7Lx42iEg6n55OQeu3vf9fJKijCydKUlyKc6dmkNl4OFhgVXiUtTdajLw+V7dgtVQH41VdfsWPHDvbv34+npyfDhg2jVatWD31Px44dSU9Pv+f5qVOnli3Shxg1ahSjRo0CICwsDFdX13Jv40Eqsq0H+ZerK1lFZny2/SyNXXOJe6qpvkMqF5Uhtx9uSeJMzk0+7+pLu6B61aK4/kdlyG91Jbm9vw2jaxH62Q6G/XyOhJdbUdPUpEz7kfw+urS8G8SnZBOfkkN8Sg77L1yhsPh2n8y6tSyI9HKgSS0jBrZojG8dG+nnrCOV6dgtVQFYUFDAK6+8QmhoKKampZv6YvPmzQ98zcXFhbS0NDQaDWlpaTg7y221x/VJbFPO5eTzf7+dwcvBkjEtvPQdUpWXkp3P+5tO0a5BbfqFuFWr4k8IfWjsYsO8ZwN5dvEBhv6QyPcDQvUdUrVUVFzC0fSrxKfksDM5m/iUbFJybnfbMjMxoqmLDf2C3QjQ2BLqYUdTFxtqW5qRnp6Gq6tM12MoSlXNTZgwoVwbjY2NZeHChUycOJGFCxfStWvXct2/ITIyMmLJgFBaf72T8SuP4lHLnC5N6+g7rCrtpVVHAXi1TX0cZWZ7IcpFnyA34lNy+GJHMi287Bnf0lvfIVV5VwoK2X0uR3uFb8+5HK7dKgbA0cqMQFdbuvtrCNTY0MzTHo9aFljLPKYGTy9HwMSJE+nduzdz586lbt26LFu2DICEhARmzpzJnDlzAGjVqhV//vkn165dw93dnblz5xITE6OPkKsEM1Njfh3VnLDPttNn8QG2j2tBiHstfYdVJf18LJ3Vxy7xQst6RDeUybaFKE+fPN2Ufeev8NqaE4S516K5l4O+Q6oylFKczrz+963cbHYm53D88lWUAmMj8HG0onMTZwI1tgS52RHkaouzTU1qmJRqzKcwIEaqmk3MFBYWRkJCQoW0lZqaWqnu5//jr5wCQj/bjpER7HuxFXUdqt58TfrMbf6tInw/2oaxkRE/Dw+naZ3qd0uksh671YHktnQyrt3E/+PfMTGCQ6+2KfVVdkPLb0FhMQl/XdEWfPEpOWRevwWATU1T/DU2BGhsCXS1JdzDjvqO1tiZm5apy4qh5baiVVR+S1sHyTXgasjD3oINo5rT6uuddJi1i4SXWmNnUUPfYVUZH2w5TUpOAbN6+lfL4k+IysDJuiarh4bR6ut4YuftY8f4KExk4AGpuXcO1sjmwMXc/wzWsLcg0tOeAI0NAa62NKtbizo25pjXKNtgGmHYpACspoLd7Vg2KJSn5+0jZvZuto+LwsxUbgH8LycvX+Oj307zZBNnegXKmbAQuhTh6cCnsb48v/IoL606ypfP+Os7pApVVFzCkbSr7Lyj4Dv392CNmibGNK1jTf9gNwJcbQl1t6PJ34M1ZISuKA9SAFZjnZu48PUzfoxZfoQ+i/ezYkiYjGR9CKUU41YcoaaJMS+2rIe9pZm+QxKi2hsX5UV8Sg5f7UyhhZcDfUPc9B2SzuTk3/p7sMbtf3vO53D978EaTn8P1ujhr9Fe3ZPBGkKX5Miq5kZHepGcdXt6mFdWH+Ozbn76DqnS+jExlS1JmbzRrj7tfBz1HY4QBsHIyIh5fQI5nJrLyGWHCHS1pWmdqr8ClFKKpMzrxCfnEH8um53J2Ry/dA0AEyPwcbKmSxNnAl1tCXK1I8jNFmfrmpjKYA1RQaQANAAfdmnCuZwCpu9IxsvBkhdby7QL/y3vRiEvrzlGE2drRjb3lBFzQlQg8xomrBsRQeAnvxM7by+HXm2DVRW78lVQWMy+81fu6r+XlV8I3B6sEaCxYWwLTwI1toTXrYV3basyD9YQojxUrd8wUSZGRkYs6hfMxdwbvLrmGJ72FnTz1+g7rEpl0oaTpOfdJO7JJrLmpRB64OlgyQ8DQ+kyZw+9Fu1n3Yhmlbo4uphbcNfI3AMXcikquT1Yw9PeghZeDgS62hKgsSG8bi00tuZlXvlECF2QAtBA1DAxZt2ICMKnb6f/dwf5baw5zTzt9R1WpZB4MZcvdiTTI0BDN3+ZPFsIfXmisTP/L7oh7248xZRNSfy/Tg31HRJwe7DG4bS8O1bWyOH8lb8Ha5ga4+tiw8BQdwI0NoS616KJizUOMlhDVHJSABoQG3NTto5tQchn23lyzh72vdSKerUN+2pXSYniuZ+OYGdeg3FRntiay3Q5QujTpE4N2X0uh3c3nSKibi06Na74pUKz7xqskc2ec1fIL7w9WMPZ2oxAjS29AjUEaGyJ8KyFu51FlbtlLYQcsQbG1c6czaObE/nFH3SYuYv9L7c26NGu8/f9xa5zOUzu1JBW3jLwQwh9MzIy4sdBoQR8/Dv9vjtA4ittcLe30Fl7SilOZVzX3srdmZzNicv/GazR0MmaWF8XAjS2BLvbEuhqh5OVmQzWEFWeFIAGyE9jy8ohYXSes5eOM3cT/0KUQfZNybp+izfWHifYzZZhzTxkElohKglb8xqsHd6MZp/voMvcvex7qVW5zWOaf6uIfdqVNXLYdcdgDduapgS42vJcC08CNLY0q2tPvdqW1JKJ9EU1JAWggerYyJlvegcw9IdDPLMggZ+HNTO4/ipv/nKCKwWFfP2MPx72VW+5PCGqMz+NLXN6BdL/+4MMX5rI4v4hZdrPhSt3D9Y4ePE/gzW87C2IqudAgMaWIFdbwjxqUce2pkGeEAvDIwWgARsSXpeUrALe3XSK51ce4eseAfoOqcLsPpfDN7vPMzDUjaeauug7HCHEffQLdWdnSjb/jj9Hi3oOjG3h9dDtC4tLOJz692CNlGziU7L568oN4PZgDb86twdrBP5d7DVyspLBGsJgSQFo4CbFNCQ5J59/x5/Dy8GSCe0a6DsknSsqLmHs8sM4W5vxXIt60nlbiErs825+JFzI5eXVxwhzt8Ptjl/X7Pxb7Lrj6t7e8/8ZrOFiU5MAjQ19Al21K2vIYA0h/kN+EwyckZERc3sHcuHKDSauO0Fdewv6BFXfpZgA/h2fQmJqHnFdGtOsbi19hyOEeAhTE2PWDGtGwMfb6L4ggRebOXNyRwY7U7L5847BGo2crenq60KAqy1BbrdX13CUwRpCPJAUgOLvL9hwIj7/g8FLEnGzM6dlvdr6Dksn0vJu8Pb6k0R62jMo1F1u/QhRBbjY1GTlkHDa/Due1zedx87clACNLR1aeBHganN7sIaDJXYyWEOIUpMCUABgVdOULWMjCf10O7Fz97H7hZY0dLbWd1jl7tU1x7lZVMzr7eqjsdPd1BJCiPLVop4Du19syfGzF2jj542LDNYQ4rHItXGh5WJTk81jmlOiFNGzdpNx7aa+QypXW5MyWXLwIoPDPHhCD5PLCiEeT6h7LTo0cKCug6UUf0I8Jr0UgNnZ2URHR+Pj40N0dDQ5OTn3bJOYmEhkZCS+vr4EBASwdOlSPURqeBq72LB6WDPSr96g48zdFPzdobqqu1VUwnM/HcbdzpyxLTwxryF/PIQQQhguvRSAcXFxdOjQgaSkJDp06EBcXNw921haWrJo0SKOHTvGr7/+yksvvcSVK1f0EK3haVO/NvP6BHEkLY+u8/ZS/PecWVXZJ7+f4WTGdSa0rU+Qm52+wxFCCCH0Si8F4OrVqxk8eDAAgwcPoxnpUwAAHHxJREFUZtWqVfds07BhQ3x8fABwdXXF2dmZjIyMCo3TkPUPdWdK50ZsOpXJqGWH9B3OY0nJzuf9Tado16A2/ULcMDKSgR9CCCEMm14GgVy6dAmNRgOARqPh8uXLD91+79693Lp1i/r169/39dmzZzN79mwA0tPTSU1NLd+AH6C6F6SDm1hx/EJt5u39C6caRbzQ3LXC2i7P3I5edRqlFCP9bbmVl0VqXrntusqq7seuPkludUvyqzuSW92qbPnVWQHYsWNH0tPT73l+6tSpj7SftLQ0Bg4cyMKFCzE2vv8Fy1GjRjFq1CgAwsLCcHWtuEKlItvSh4WDXMmcs4ePdqbhW9eFgWEeFdZ2eeT252PpbDyTywst69GjedNyW0+0Oqjux64+SW51S/KrO5Jb3apM+dVZAbh58+YHvubi4kJaWhoajYa0tDScne8/IjMvL48uXbowZcoUmjdvrqtQxUOYGBuxcmg4Lb74gxE/HsKjlgVtGzjqO6xSyb9VxAsrj+LtYMnoyLpS/AkhhBB/08tfxNjYWBYuXAjAwoUL6dq16z3b3Lp1i+7duzNo0CB69epV0SGKO1jUMGHT6Oa42NSk2/x9nLh0Vd8hlcoHW06TklPAxPYNaFrHVt/hCCGEEJWGXgrAiRMnsmnTJnx8fNi0aRMTJ04EICEhgREjRgDw448/sn37dhYsWEBQUBBBQUEkJibqI1wBOFrXZMuYSIyMIHrWbi7l3dB3SA918vI1PvrtNF2aONMzUKPvcIQQQohKRS+DQGrXrs2WLVvueT4sLIw5c+YAMGDAAAYMGFDRoYmH8HGyZt2ICNrP2EX7mbvY+2KrSrmwulKKcSuOUNPEmBda1sPe0kzfIQkhhBCVinSKEo+khZcD3/YL5s/L13hq7l6Kikv0HdI9liamsiUpk3FRXrTzqRr9FYUQQoiKJAWgeGQ9A12Z9lRTtp3JYsgPiShVeSaKzi0o5OXVx2jibM3I5p7UMJFDXAghhPhvle/+nagSXmlbn+TsfL7amUI9B0ve79xY3yEBMGnDSS5dvcn/dWlCfUcrfYcjhBBCVEpSAIoy+7ybH+dyCpiyOQkvewuGN/fUazyJF3P58o9kegRo6OZfR6+xCCGEEJWZ3B8TZWZsbMSPg0IJc7dj7IojbDz58BVddKmkRPHcT0ewM6/BuChPbM1r6C0WIYQQorKTAlA8FvMaJmwY3Rw3W3N6LtzP4dRcvcQxf99f7DqXw4ut6tHKWwZ+CCGEEA8jBaB4bA6WZvz2XAvMTIyImb2Hi1cKKrT9zGs3eX3tcYLdbBnWzAMTY6MKbV8IIYSoaqQAFOXCy8GS9SMjyL1RSIeZu7h6o6jC2n7zlz/JLSjkjXYN8LC3rLB2hRBCiKpKCkBRbsLr2rNkQCinM6/T+ZvdFFbAHIG7UrKZs+c8/ULceKqpi87bE0IIIaoDKQBFuerqV4fp3fzYmZJD/+8O6HSOwKLiEp776QjO1mY816JepVyVRAghhKiMpAAU5W58y3q80sabZYfSeGPtCZ218+/4FBJT83i1TX2a1a2ls3aEEEKI6kYumQid+PjpppzLyWfatjN4OVjwXFS9ct1/Wt4N3l5/khae9gwMdcNYBn4IIYQQpSZXAIVOGBkZ8X3/UCI97Xlh1TF+PpZervt/dc1xbhYVM6FdfTR2FuW6byGEEKK6kwJQ6IyZqTHrR0ZQz96Cvt8eYP+FK+Wy3y2nMlhy8CKDwzx4orFzuexTCCGEMCRSAAqdsrOowdaxLbAyM+HJb/ZwPjv/sfZ3s6iYcSuO4G5nztgWnpjXMCmnSIUQQgjDIQWg0DkPews2jGzO9VvFtJ+5i9yCwjLv65NtZzmZcZ0JbesT5GZXjlEKIYQQhkMKQFEhgtztWDYolJScAjrN2s2tokefIzAlO58pm0/RvkFt+oW4YWQkAz+EEEKIstBLAZidnU10dDQ+Pj5ER0eTk5Nzzzbnzp0jNDSUoKAgfH19mTlzph4iFeWpcxMX/v2MH3v/ukLvRQmPPEfgi6uOAvBKm/o4WtfURYhCCCGEQdBLARgXF0eHDh1ISkqiQ4cOxMXF3bONRqMhPj6exMRE9uzZQ1xcHKmpqXqIVpSnUZFevNWhAauPXeLl1cdK/b41R9NZc+wSIyM8iW7opMMIhRBCiOpPLwXg6tWrGTx4MACDBw9m1apV92xjZmZGzZq3r/LcvHmTkhLdLysmKsaUzo3pG+zK5zuSmb79zP/cPv9WES+sOoq3gyWjI+tiZio9F4QQQojHoZeJoC9duoRGowFuX+m7fPnyfbf766+/6NKlC6dPn2batGm4urred7vZs2cze/ZsANLT0yvsSmFGRkaFtFMdfdimDsmXc3l1zXFs1Q2e8LG/6/U7cxv3x0XO5RTwabQ7tUqukZp6raLDrXbk2NUdya1uSX51R3KrW5UtvzorADt27Eh6+r2T/06dOrXU+/Dw8ODw4cOkpqbSrVs3evbsiYuLyz3bjRo1ilGjRgEQFhb2wEJRFyqyrepm43MuNPt8B+PXp7DV05XmXncXga6urvx56SozEy7RpYkzQ1r7Ym9ppqdoqx85dnVHcqtbkl/dkdzqVmXKr84KwM2bNz/wNRcXF9LS0tBoNKSlpeHs/PDJfF1dXfH19WXHjh307NmzvEMVemJjbsrWsZGEfLqdp+buYe+LrfB2tNK+rpRi/MqjmJua8ELLelL8CSGEEOVEL52pYmNjWbhwIQALFy6ka9eu92xz4cIFCgoKAMjJyWHnzp00atSoQuMUuqexNWfz6ObcLCqhw8xdZOff0r62NDGVLUmZjGvhSXsfRz1GKYQQQlQveikAJ06cyKZNm/Dx8WHTpk1MnDgRgISEBEaMGAHAiRMniIiIIDAwkDZt2vDaa6/h7++vj3CFjvlqbFk1NJyLeTeInrmLm0XF5N0s5uXVx2jqYs2I5p6YmsjADyGEEKK86GUQSO3atdmyZcs9z4eFhTFnzhwAoqOjOXz4cEWHJvSkQ0MnvukVyJAfEuk+fx8aC7h09SYfPdWE+nfcFhZCCCHE49NLASjE/QwO9yAlO5/JG08B0DNAQzc/jZ6jEkIIIaofKQBFpfJOp4ak5d1k7bFUxkV5YmMuh6gQQghR3uSvq6hUjIyMmNkrgFNhdjTwlIEfQgghhC5Iz3pRKVnXrIGxsZG+wxBCCCGqJSkAhRBCCCEMjBSAQgghhBAGRgpAIYQQQggDIwWgEEIIIYSBkQJQCCGEEMLAGCmllL6DKE+Ojo54eXlVSFsZGRk4OTlVSFuGRnKrW5Jf3ZHc6pbkV3ckt7pVUflNSUkhMzPzf25X7QrAihQWFkZCQoK+w6iWJLe6JfnVHcmtbkl+dUdyq1uVLb9yC1gIIYQQwsBIASiEEEIIYWBMJk+ePFnfQVRloaGh+g6h2pLc6pbkV3ckt7ol+dUdya1uVab8Sh9AIYQQQggDI7eAhRBCCCEMjBSAQgghhBAGxmALwKlTp+Lr60tAQABBQUHs2bPnsfc5efJkPv7443KIruoyMjJi4MCB2sdFRUU4OTnx1FNPlcv+DTHHWVlZBAUFERQURJ06dXBzc9M+vnXrVrm317JlSxITE8t9v/rw8ssvM336dO3jmJgYRowYoX386quv8umnn5ZqX7o+9hYsWMD48eN1tv+K8qDjtVatWjRt2lTn7VeXPD4OExMT7c8gKCiIlJSUe7ZJTU2lZ8+e931/27ZtK9V0JfrwKDXCggULSE1Nfew2KzrvphXWUiWya9cu1q5dy4EDB6hZsyaZmZk6+UNqiKysrDh69CgFBQVYWFiwadMm3Nzc9B1WlVa7dm1tQTZ58mSsra157bXX9BxV1dCiRQuWLVvGSy+9RElJCZmZmeTl5Wlfj4+Pv6tAFI/vQcdrSkrKY50IFhUVYWpqkH+yHpmFhcVDT+KKiopwdXVl+fLlFRhV1fGoNcKCBQvw8/PD1dW11G1UhuPZIK8ApqWl4ejoSM2aNYHbq4e4urri5eWlnT07ISGBtm3bAre/xIYNG0bbtm3x9vbmiy++0O5r6tSpNGrUiI4dO3Ly5Ent89988w3h4eEEBgbSo0cP8vPzuXr1KvXq1aOwsBCAvLw8vLy8tI+ri86dO7Nu3ToAlixZQt++fbWvZWdn061bNwICAmjevDmHDx8GJMdlcfr0aYKCgrSP4+LimDJlCgBJSUnExMQQGhpK69atOXXqFAA//PADfn5+BAYG0q5dOwDy8/Pp1asXAQEBPPvss9y4cUO7z1GjRhEWFoavry/vvfceABs2bKBXr17abdavX0/v3r11/nnLIioqivj4eACOHTuGn58fNjY25OTkcPPmTU6cOEFwcDDTpk0jPDycgIAAJk2apH3/g469tm3b8sYbb9CsWTMaNmzIjh07ACguLmbChAnafc2aNQu4/Z3TunVrgoKC8PPz024/f/58GjZsSJs2bdi5c6d2/z///DMREREEBwfTsWNHLl26RElJCT4+PmRkZABQUlJCgwYNSjXjf2VRXFzMyJEj8fX1pVOnThQUFAB3X/nIzMzUrua0YMECevXqxdNPP02nTp0kj4/hv3OZkpKCn58fAAUFBTz77LMEBATQp08f7c8FYOzYsdrvgH9+N7Zs2UL37t2122zatIlnnnmmYj+QDj2oRnjvvfcIDw/Hz8+PUaNGoZRi+fLlJCQk0L9/f4KCgigoKHhoLTFq1Cg6derEoEGD9J93ZYCuXr2qAgMDlY+Pjxo7dqzatm2bUkopT09PlZGRoZRSat++fapNmzZKKaUmTZqkIiMj1Y0bN1RGRoZycHBQt27dUgkJCcrPz09dv35d5ebmqvr166tp06YppZTKzMzUtvevf/1LffHFF0oppYYMGaJWrlyplFJq1qxZ6pVXXqmoj10hrKys1KFDh1SPHj1UQUGBCgwMVL/99pvq0qWLUkqp8ePHq8mTJyullNqyZYsKDAxUSkmOS2vSpEnaz5+UlKTNn1JKffjhh+r9999XSinVtm1bdfr0aaWUUn/88YeKjo5WSinVuHFjlZ6erpRSKicnRyml1P/93/+pkSNHKqWUOnDggDI2NlYHDx5USimVlZWllFKqsLBQtWzZUh07dkwVFxerhg0bavPfq1cv9csvv+j0cz8OT09Pde7cOTVz5kw1Y8YM9fbbb6t169apP/74Q7Vq1Upt2LBBjRw5UpWUlKji4mLVpUsX9fvvvz/02GvTpo32uFq3bp3q0KGDUur28fbPz+DGjRsqNDRUnT17Vn388cdqypQpSimlioqKVF5enkpNTVUeHh7q8uXL6ubNm6pFixZq3LhxSimlsrOzVUlJiVJKqW+++Ubb1uTJk9Vnn32mlFJqw4YN6plnnqmgLJbNncdrcnKyMjEx0R5bvXr1UosXL1ZK3c7nvn37lFJKZWRkKE9PT6WUUvPnz1dubm7a49BQ8/iojI2NVWBgoAoMDFTdunVTSt2by+TkZOXr66uUUuqTTz5RQ4cOVUopdejQIWViYqL9efyzfVFRkWrTpo06dOiQKikpUY0aNVKXL19WSinVt29ftWbNmgr9jLr0oBrhn1wopdSAAQO0n/nO41eph9cSISEhKj8/Xyml/7wb5BVAa2tr9u/fz+zZs3FycqJPnz4sWLDgoe/p0qULNWvWxNHREWdnZy5dusSOHTvo3r07lpaW2NraEhsbq93+6NGjtGrVCn9/f7777juOHTsGwIgRI5g/fz5w+6x16NChOvuc+hIQEEBKSgpLlizhySefvOu1P/74Q9tHsH379mRlZZGbmwtIjsvLlStX2L17Nz169CAoKIhx48Zp+6dERUUxaNAg5syZQ0lJCQDbt29nwIABAAQHB+Pr66vd15IlSwgJCSEkJIQTJ05w/PhxjI2N6devH99//z3Z2dns37+fTp06VfwHLaV/rgLGx8cTGRlJZGSk9nGLFi3YuHEjGzduJDg4mJCQEP7880+SkpIeeuwB2jPv0NBQbR+rjRs3smjRIoKCgoiIiCArK4ukpCTCw8OZP38+kydP5siRI9jY2LBnzx7atm2Lk5MTZmZm9OnTR7vvCxcuEBMTg7+/P9OmTdMe28OGDWPRokUAzJs3r8od2/Xq1dNetb4zbw8THR2Ng4MDgOSxlP65BZyYmMjKlSu1z9+Zyzvd+R0QEBBAQECA9rUff/yRkJAQgoODOXbsGMePH9f29f7222+5cuUKu3btonPnzrr/YBXkQTXCb7/9RkREBP7+/mzdulV7PD2K2NhYLCwsAP3n3WA7VJiYmNC2bVvatm2Lv78/CxcuxNTUVPtH8c7bYID2UvA/7y0qKgJuD3q4nyFDhrBq1SoCAwNZsGAB27ZtA27/MUpJSeH333+nuLhYewm+uomNjeW1115j27ZtZGVlaZ9X95l28p8cSo4fzZ3HK9w+Zk1NTVFK4ejoeN8+QN988w179uxh7dq1BAYGam/B3y/HSUlJfP755+zdu5datWoxYMAA7e/FsGHD6NGjBwB9+vTBxMREFx+xXLRo0YL4+HiOHDmCn58fHh4efPLJJ9ja2jJs2DC2bdvGm2++yejRo+963/Tp0x947MF/jtc7j1WlFF9++SUxMTH3bL99+3bWrVvHwIEDmTBhAra2tg/c//PPP88rr7xCbGws27Zt45/5+j08PHBxcWHr1q3s2bOH7777riwp0Zv//h3/55bXw757raystP9v3bq15PEx3JnL/3a/HCYnJ/Pxxx+zb98+7O3tGTJkiPbnM3ToUJ5++mnMzc3p1auX3vuzlbf/rhFmzZrF4cOHSUhIwMPDg8mTJ99zrP6jtMcz6DfvBnkF8OTJkyQlJWkfJyYm4unpiZeXF/v37wfgp59++p/7ad26NStXrqSgoICrV6/y888/a1+7evUqGo2GwsLCe75cBg0aRN++favdWeedhg0bxjvvvIO/v/9dz7du3Vqbj23btuHo6Iitre0D9yM5frA6deqQmppKTk4ON27c0Pa7tLe3R6PRaM/8S0pKOHToEABnz56lefPmvP/++9jb23Px4sW7fiaHDh3SntXm5eVhY2ODra0taWlpbNiwQdu2h4cHjo6OxMXFMWTIkAr81I8uKiqKtWvX4uDggImJCQ4ODtqz58jISGJiYpg3bx7Xrl0D4OLFi1y+fPmhx96DxMTEMGPGDG2f01OnTnH9+nXOnTuHs7MzI0eOZPjw4Rw4cICIiAjtCVJhYSHLli3T7ic3N1c7eGrhwoV3tTFixAgGDBhA7969K3Xh/Sju/O592MAEyaNu3PkdcPToUe2JYV5eHlZWVtjZ2XHp0iXWr1+vfY+rqyuurq5MmTKl0n8HPKr71QiNGjUCbvcHvHbt2l3HqY2NDVevXtU+Lm0toe+8V6+SvZSuXbvG888/z5UrVzA1NaVBgwbMnj2bEydOMHz4cD744AMiIiL+535CQkLo06cPQUFBeHp60qpVK+1r77//PhEREXh6euLv73/XwdG/f3/efvvtuwZHVDfu7u68+OKL9zw/efJkhg4dSkBAAJaWlvd8Kf83yfGDmZub89ZbbxEeHo63t/ddU2z88MMPjB07lsmTJ3Pr1i0GDBhAYGAgL7/8MsnJySil6NSpE35+fnh7ezN48GACAgIICQkhLCwMuJ37pk2bareJioq6q/1+/fqRl5dHw4YNK/RzPyp/f38yMzPp16/fXc9du3YNR0dHOnXqxIkTJ4iMjARu3/759ttvH3rsPciIESNISUkhJCQEpRROTk6sWrWKbdu2MW3aNGrUqIG1tTWLFi1Co9EwefJkIiMj0Wg0hISEUFxcDNz+PenVqxdubm40b96c5ORkbRuxsbEMHTq0Wp3cvPbaa/Tu3ZvFixfTvn37B24nedSNsWPHar+Xg4KCaNasGQCBgYHabiH3+w7o378/GRkZFTK9T0V6UI1Qq1Yt/P398fLyIjw8XLv9kCFDGDNmDBYWFuzatYtJkyaVqpbQd95lKTg9WL58OatXr2bx4sX6DqXakhzr3pgxY4iMjGTw4MH6DsWgJCQk8PLLL2tHwIqykTw+vvHjxxMcHMzw4cP1HYpBKa+8G+QVQH16/vnnWb9+Pb/88ou+Q6m2JMe6FxQUhL29/V3T9Qjdi4uLY8aMGQbTZ01XJI+PLzQ0FCsrKz755BN9h2JQyjPvcgVQCCGEEMLAGOQgECGEEEIIQyYFoBBCCCGEgZECUAghhBDCwEgBKISolkxMTAgKCsLX15fAwEA+/fTTuybO1oUJEybg6+vLhAkTdNrOneu4CiFEWcgoYCFEtfTPclgAly9fpl+/fuTm5vLuu+/qrM1Zs2aRkZFx14oXQghRGckVQCFEtefs7Mzs2bP56quvUEqRkpJCq1attOscx8fHAzBw4EBWr16tfV///v1Zs2bNXftSSjFhwgT8/Pzw9/dn6dKlwO2Jha9fv05ERIT2uX/4+/tz5coVlFLUrl1buw7twIED2bx5M8XFxUyYMIHw8HACAgKYNWuW9r3Tpk3TPj9p0qR7PtvZs2cJDg5m37595ZMsIYRBkCuAQgiD4O3tTUlJCZcvX8bZ2ZlNmzZhbm5OUlISffv2JSEhgREjRvDZZ5/RtWtXcnNziY+Pv2e1mhUrVpCYmMihQ4fIzMwkPDyc1q1bs2bNGqytre+7BnNUVBQ7d+7E09MTb29vduzYwaBBg9i9ezczZsxg7ty52NnZsW/fPm7evElUVBSdOnUiKSmJpKQk9u7di1KK2NhYtm/fTt26dYHbS1Y9++yzzJ8/n6CgoArJoxCiepACUAhhMP6Z9rSwsJDx48eTmJiIiYkJp06dAqBNmzaMGzeOy5cvs2LFCnr06HHPYut//PEHffv2xcTEBBcXF9q0acO+ffuIjY19YLutWrVi+/bteHp6MnbsWGbPns3FixdxcHDA2tqajRs3cvjwYe36orm5uSQlJbFx40Y2btxIcHAwcHuJqqSkJOrWrUtGRgZdu3blp59+wtfXVxfpEkJUY1IACiEMwtmzZzExMcHZ2Zl3330XFxcXDh06RElJCebm5trtBg4cyHfffccPP/zAvHnz7tlPWebOb926NV9//TXnz59n6tSprFy5kuXLl2vXF1ZK8eWXXxITE3PX+zZs2MCbb77J6NGj73o+JSUFOzs7PDw82LlzpxSAQohHJn0AhRDVXkZGBmPGjGH8+PEYGRmRm5uLRqPB2NiYxYsXU1xcrN12yJAhTJ8+HeC+hVXr1q1ZunQpxcXFZGRksH37du0i7g/i4eFBZmYmSUlJeHt707JlSz7++GNtARgTE8OMGTMoLCwE4NSpU1y/fp2YmBjmzZvHtWvXALh48SKXL18GwMzMjFWrVrFo0SK+//77x0+SEMKgyBVAIUS1VFBQQFBQEIWFhZiamjJw4EBeeeUVAJ577jl69OjBsmXLaNeuHVZWVtr3ubi40KRJE7p163bf/Xbv3p1du3YRGBiIkZERH330EXXq1Pmf8URERGgLzVatWvHmm2/SsmVLAEaMGEFKSgohISEopXBycmLVqlV06tSJEydOEBkZCYC1tTXffvstJiYmAFhZWbF27Vqio6OxsrKia9euZU+YEMKgyFrAQghxh/z8fPz9/Tlw4AB2dnb6DkcIIXRCbgELIcTfNm/eTOPGjXn++eel+BNCVGtyBVAIIYQQwsDIFUAhhBBCCAMjBaAQQgghhIGRAlAIIYQQwsBIASiEEEIIYWCkABRCCCGEMDD/H5Uhvq3Y1A0dAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 648x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "prophet_model.plot_components(forecast_data);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Scaling back results to compare to reality"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [],
   "source": [
    "forecast_data['y_or'] = std_scale.inverse_transform(forecast_data['yhat'])\n",
    "forecast_data['y_l_or'] = std_scale.inverse_transform(forecast_data['yhat_lower'])\n",
    "forecast_data['y_h_or'] = std_scale.inverse_transform(forecast_data['yhat_upper'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "x and y must have same first dimension, but have shapes (112,) and (49,)",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-200-c47787b5484b>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      3\u001b[0m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mplot\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mforecast_data\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'ds'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mforecast_data\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'y_l_or'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'b--'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      4\u001b[0m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mplot\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mforecast_data\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'ds'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mforecast_data\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'y_h_or'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'b--'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 5\u001b[1;33m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mplot\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mforecast_data\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'ds'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdf_test_mean\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'y_ori'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      6\u001b[0m \u001b[0max\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mset_xlabel\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"Testing Days\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      7\u001b[0m \u001b[0max\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mset_ylabel\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"Daily mean delay\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m;\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\matplotlib\\pyplot.py\u001b[0m in \u001b[0;36mplot\u001b[1;34m(*args, **kwargs)\u001b[0m\n\u001b[0;32m   3356\u001b[0m                       mplDeprecation)\n\u001b[0;32m   3357\u001b[0m     \u001b[1;32mtry\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 3358\u001b[1;33m         \u001b[0mret\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0max\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mplot\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m*\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   3359\u001b[0m     \u001b[1;32mfinally\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3360\u001b[0m         \u001b[0max\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_hold\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mwashold\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\matplotlib\\__init__.py\u001b[0m in \u001b[0;36minner\u001b[1;34m(ax, *args, **kwargs)\u001b[0m\n\u001b[0;32m   1853\u001b[0m                         \u001b[1;34m\"the Matplotlib list!)\"\u001b[0m \u001b[1;33m%\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mlabel_namer\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfunc\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m__name__\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1854\u001b[0m                         RuntimeWarning, stacklevel=2)\n\u001b[1;32m-> 1855\u001b[1;33m             \u001b[1;32mreturn\u001b[0m \u001b[0mfunc\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0max\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m*\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1856\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1857\u001b[0m         inner.__doc__ = _add_data_doc(inner.__doc__,\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\matplotlib\\axes\\_axes.py\u001b[0m in \u001b[0;36mplot\u001b[1;34m(self, *args, **kwargs)\u001b[0m\n\u001b[0;32m   1525\u001b[0m         \u001b[0mkwargs\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mcbook\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnormalize_kwargs\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0m_alias_map\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1526\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1527\u001b[1;33m         \u001b[1;32mfor\u001b[0m \u001b[0mline\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_get_lines\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m*\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1528\u001b[0m             \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0madd_line\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mline\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1529\u001b[0m             \u001b[0mlines\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mline\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\matplotlib\\axes\\_base.py\u001b[0m in \u001b[0;36m_grab_next_args\u001b[1;34m(self, *args, **kwargs)\u001b[0m\n\u001b[0;32m    404\u001b[0m                 \u001b[0mthis\u001b[0m \u001b[1;33m+=\u001b[0m \u001b[0margs\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    405\u001b[0m                 \u001b[0margs\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0margs\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 406\u001b[1;33m             \u001b[1;32mfor\u001b[0m \u001b[0mseg\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_plot_args\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mthis\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mkwargs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    407\u001b[0m                 \u001b[1;32myield\u001b[0m \u001b[0mseg\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    408\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\matplotlib\\axes\\_base.py\u001b[0m in \u001b[0;36m_plot_args\u001b[1;34m(self, tup, kwargs)\u001b[0m\n\u001b[0;32m    381\u001b[0m             \u001b[0mx\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mindex_of\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mtup\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m-\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    382\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 383\u001b[1;33m         \u001b[0mx\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_xy_from_xy\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mx\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    384\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    385\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcommand\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;34m'plot'\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\matplotlib\\axes\\_base.py\u001b[0m in \u001b[0;36m_xy_from_xy\u001b[1;34m(self, x, y)\u001b[0m\n\u001b[0;32m    240\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mx\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m!=\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    241\u001b[0m             raise ValueError(\"x and y must have same first dimension, but \"\n\u001b[1;32m--> 242\u001b[1;33m                              \"have shapes {} and {}\".format(x.shape, y.shape))\n\u001b[0m\u001b[0;32m    243\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mx\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mndim\u001b[0m \u001b[1;33m>\u001b[0m \u001b[1;36m2\u001b[0m \u001b[1;32mor\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mndim\u001b[0m \u001b[1;33m>\u001b[0m \u001b[1;36m2\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    244\u001b[0m             raise ValueError(\"x and y can be no greater than 2-D, but have \"\n",
      "\u001b[1;31mValueError\u001b[0m: x and y must have same first dimension, but have shapes (112,) and (49,)"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(20,5))\n",
    "plt.plot(forecast_data['ds'], forecast_data['y_or'])\n",
    "plt.plot(forecast_data['ds'], forecast_data['y_l_or'], 'b--')\n",
    "plt.plot(forecast_data['ds'], forecast_data['y_h_or'], 'b--')\n",
    "plt.plot(forecast_data['ds'], df_test_mean['y_ori'])\n",
    "ax.set_xlabel(\"Testing Days\")\n",
    "ax.set_ylabel(\"Daily mean delay\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "rmse_test = math.sqrt(metrics.mean_squared_error(df_test_mean['y_ori'].values, forecast_data['y_or'].values))\n",
    "mae_test = metrics.mean_absolute_error(df_test_mean['y_ori'].values, forecast_data['y_or'].values)\n",
    "print(\"RMSE on testing dataset : %.2f\" % rmse_test)\n",
    "print(\"MAE on testing dataset : %.2f\" % mae_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Found input variables with inconsistent numbers of samples: [112, 231]",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-201-41ee5a05b8d4>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[0mtrain_predict\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mstd_scale\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minverse_transform\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprophet_model\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mpast_data\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'yhat'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m \u001b[0mrmse_train\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mmath\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msqrt\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmetrics\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmean_squared_error\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdf_train_mean\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'y'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtrain_predict\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      3\u001b[0m \u001b[0mmae_train\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mmetrics\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmean_absolute_error\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdf_train_mean\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'y'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtrain_predict\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      4\u001b[0m \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"RMSE on training dataset : %.2f\"\u001b[0m \u001b[1;33m%\u001b[0m \u001b[0mrmse_train\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      5\u001b[0m \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"MAE on training dataset : %.2f\"\u001b[0m \u001b[1;33m%\u001b[0m \u001b[0mmae_train\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\regression.py\u001b[0m in \u001b[0;36mmean_squared_error\u001b[1;34m(y_true, y_pred, sample_weight, multioutput)\u001b[0m\n\u001b[0;32m    236\u001b[0m     \"\"\"\n\u001b[0;32m    237\u001b[0m     y_type, y_true, y_pred, multioutput = _check_reg_targets(\n\u001b[1;32m--> 238\u001b[1;33m         y_true, y_pred, multioutput)\n\u001b[0m\u001b[0;32m    239\u001b[0m     output_errors = np.average((y_true - y_pred) ** 2, axis=0,\n\u001b[0;32m    240\u001b[0m                                weights=sample_weight)\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\metrics\\regression.py\u001b[0m in \u001b[0;36m_check_reg_targets\u001b[1;34m(y_true, y_pred, multioutput)\u001b[0m\n\u001b[0;32m     73\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     74\u001b[0m     \"\"\"\n\u001b[1;32m---> 75\u001b[1;33m     \u001b[0mcheck_consistent_length\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0my_true\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my_pred\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     76\u001b[0m     \u001b[0my_true\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mcheck_array\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0my_true\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mensure_2d\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mFalse\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     77\u001b[0m     \u001b[0my_pred\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mcheck_array\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0my_pred\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mensure_2d\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mFalse\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\utils\\validation.py\u001b[0m in \u001b[0;36mcheck_consistent_length\u001b[1;34m(*arrays)\u001b[0m\n\u001b[0;32m    202\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0muniques\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m>\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    203\u001b[0m         raise ValueError(\"Found input variables with inconsistent numbers of\"\n\u001b[1;32m--> 204\u001b[1;33m                          \" samples: %r\" % [int(l) for l in lengths])\n\u001b[0m\u001b[0;32m    205\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    206\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mValueError\u001b[0m: Found input variables with inconsistent numbers of samples: [112, 231]"
     ]
    }
   ],
   "source": [
    "train_predict = std_scale.inverse_transform(prophet_model.predict(past_data)['yhat'])\n",
    "rmse_train = math.sqrt(metrics.mean_squared_error(df_train_mean['y'].values, train_predict))\n",
    "mae_train = metrics.mean_absolute_error(df_train_mean['y'].values, train_predict)\n",
    "print(\"RMSE on training dataset : %.2f\" % rmse_train)\n",
    "print(\"MAE on training dataset : %.2f\" % mae_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig, ax = plt.subplots(figsize=(20,5))\n",
    "plt.plot(past_data['ds'], train_predict)\n",
    "plt.plot(past_data['ds'], df_train_mean['y_ori'])\n",
    "ax.set_xlabel(\"Training Days\")\n",
    "ax.set_ylabel(\"Daily mean delay\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# del X, X_train_std, ax1, ax2, data, data_reg, label_airlines, oh_airlines, oh_dep_airport, pred_df, predict_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "whos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}