SCI16. Forward Elimination and Modified Pipeline with parameterized input feature list and response.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "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>Time</th>\n",
       "      <th>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>V9</th>\n",
       "      <th>...</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>254969</td>\n",
       "      <td>157003.0</td>\n",
       "      <td>0.121685</td>\n",
       "      <td>0.990536</td>\n",
       "      <td>-0.449069</td>\n",
       "      <td>-0.555860</td>\n",
       "      <td>1.004292</td>\n",
       "      <td>-0.798871</td>\n",
       "      <td>0.955062</td>\n",
       "      <td>-0.105045</td>\n",
       "      <td>-0.048104</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.343563</td>\n",
       "      <td>-0.852037</td>\n",
       "      <td>0.093501</td>\n",
       "      <td>0.474447</td>\n",
       "      <td>-0.401226</td>\n",
       "      <td>0.118758</td>\n",
       "      <td>0.220506</td>\n",
       "      <td>0.084387</td>\n",
       "      <td>4.99</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>120088</td>\n",
       "      <td>75706.0</td>\n",
       "      <td>-0.644431</td>\n",
       "      <td>0.616995</td>\n",
       "      <td>-0.034061</td>\n",
       "      <td>-1.542783</td>\n",
       "      <td>2.884229</td>\n",
       "      <td>3.102120</td>\n",
       "      <td>0.222305</td>\n",
       "      <td>0.897864</td>\n",
       "      <td>-0.643099</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.239270</td>\n",
       "      <td>-1.000861</td>\n",
       "      <td>-0.075247</td>\n",
       "      <td>0.932284</td>\n",
       "      <td>-0.001419</td>\n",
       "      <td>-0.044064</td>\n",
       "      <td>0.037365</td>\n",
       "      <td>0.113453</td>\n",
       "      <td>9.21</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>69784</td>\n",
       "      <td>53581.0</td>\n",
       "      <td>1.083448</td>\n",
       "      <td>0.237004</td>\n",
       "      <td>0.547034</td>\n",
       "      <td>1.007855</td>\n",
       "      <td>-0.168149</td>\n",
       "      <td>-0.187504</td>\n",
       "      <td>0.038606</td>\n",
       "      <td>0.017704</td>\n",
       "      <td>-0.383930</td>\n",
       "      <td>...</td>\n",
       "      <td>0.066481</td>\n",
       "      <td>0.209556</td>\n",
       "      <td>-0.012882</td>\n",
       "      <td>0.235400</td>\n",
       "      <td>0.437695</td>\n",
       "      <td>-0.445046</td>\n",
       "      <td>0.037423</td>\n",
       "      <td>0.020770</td>\n",
       "      <td>29.94</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>127236</td>\n",
       "      <td>78257.0</td>\n",
       "      <td>-1.005030</td>\n",
       "      <td>1.085609</td>\n",
       "      <td>-0.287963</td>\n",
       "      <td>-1.313881</td>\n",
       "      <td>2.402693</td>\n",
       "      <td>3.304870</td>\n",
       "      <td>0.078386</td>\n",
       "      <td>0.918765</td>\n",
       "      <td>-0.063993</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.341013</td>\n",
       "      <td>-0.936019</td>\n",
       "      <td>0.057978</td>\n",
       "      <td>0.973877</td>\n",
       "      <td>0.079936</td>\n",
       "      <td>0.087449</td>\n",
       "      <td>0.357701</td>\n",
       "      <td>0.234344</td>\n",
       "      <td>0.89</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>163403</td>\n",
       "      <td>115907.0</td>\n",
       "      <td>1.940847</td>\n",
       "      <td>-0.863129</td>\n",
       "      <td>-0.308239</td>\n",
       "      <td>0.667133</td>\n",
       "      <td>-1.108577</td>\n",
       "      <td>-0.229832</td>\n",
       "      <td>-0.905282</td>\n",
       "      <td>0.054437</td>\n",
       "      <td>-0.286118</td>\n",
       "      <td>...</td>\n",
       "      <td>0.062379</td>\n",
       "      <td>0.724249</td>\n",
       "      <td>0.060358</td>\n",
       "      <td>0.039766</td>\n",
       "      <td>-0.054572</td>\n",
       "      <td>-0.402665</td>\n",
       "      <td>0.062921</td>\n",
       "      <td>-0.034682</td>\n",
       "      <td>50.00</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Time        V1        V2        V3        V4        V5        V6  \\\n",
       "254969  157003.0  0.121685  0.990536 -0.449069 -0.555860  1.004292 -0.798871   \n",
       "120088   75706.0 -0.644431  0.616995 -0.034061 -1.542783  2.884229  3.102120   \n",
       "69784    53581.0  1.083448  0.237004  0.547034  1.007855 -0.168149 -0.187504   \n",
       "127236   78257.0 -1.005030  1.085609 -0.287963 -1.313881  2.402693  3.304870   \n",
       "163403  115907.0  1.940847 -0.863129 -0.308239  0.667133 -1.108577 -0.229832   \n",
       "\n",
       "              V7        V8        V9  ...       V21       V22       V23  \\\n",
       "254969  0.955062 -0.105045 -0.048104  ... -0.343563 -0.852037  0.093501   \n",
       "120088  0.222305  0.897864 -0.643099  ... -0.239270 -1.000861 -0.075247   \n",
       "69784   0.038606  0.017704 -0.383930  ...  0.066481  0.209556 -0.012882   \n",
       "127236  0.078386  0.918765 -0.063993  ... -0.341013 -0.936019  0.057978   \n",
       "163403 -0.905282  0.054437 -0.286118  ...  0.062379  0.724249  0.060358   \n",
       "\n",
       "             V24       V25       V26       V27       V28  Amount  Class  \n",
       "254969  0.474447 -0.401226  0.118758  0.220506  0.084387    4.99      0  \n",
       "120088  0.932284 -0.001419 -0.044064  0.037365  0.113453    9.21      0  \n",
       "69784   0.235400  0.437695 -0.445046  0.037423  0.020770   29.94      0  \n",
       "127236  0.973877  0.079936  0.087449  0.357701  0.234344    0.89      0  \n",
       "163403  0.039766 -0.054572 -0.402665  0.062921 -0.034682   50.00      0  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# importing packages\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import scipy.stats as stats\n",
    "import sklearn\n",
    "import imblearn\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "plt.style.use('ggplot')\n",
    "\n",
    "# reading in CSV\n",
    "df = pd.read_csv('creditcard.csv')\n",
    "df.sample(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Remove irrelevant features like time before training model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('Time', axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Full code for machine learning pipeline\n",
    "### Note: SCI17 Jupyter Notebook and sklearn_classification_pipeline.py has the more updated version\n",
    "### Pipeline allows one to select variable list to use in model, to standardize data, ratio of data for testing in train-test\n",
    "### Type of sampling - naive, smote, adasyn (with options to add more)\n",
    "### Type of model\n",
    "### Custom text for decision matrix shown as heatmap\n",
    "### Whether to enable or disable CV (this is changed to a custom built stratified CV in SCI17 and py code)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# importing packages\n",
    "import math\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import scipy.stats as stats\n",
    "import sklearn\n",
    "import imblearn\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "plt.style.use('ggplot')\n",
    "\n",
    "from imblearn.over_sampling import RandomOverSampler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.svm import LinearSVC\n",
    "import xgboost as xgb \n",
    "from sklearn import metrics\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from imblearn.over_sampling import ADASYN\n",
    "import math\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.utils import shuffle\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.calibration import CalibratedClassifierCV\n",
    "from collections import Counter\n",
    "\n",
    "class modelpipeline:\n",
    "    def __init__(self):\n",
    "        pass\n",
    "    \n",
    "    def run_model(self, df, varlist, response, testratio, standardize, sampletype, modelname, text, CV):\n",
    "        # Align field orders in df (including response)\n",
    "        df = df[varlist]\n",
    "\n",
    "        if sampletype == 'smote':\n",
    "            X_train, X_test, y_train, y_test = sampling.smote_oversample(df, testratio, response)\n",
    "        elif sampletype == 'adasyn':\n",
    "            X_train, X_test, y_train, y_test = sampling.adasyn_oversample(df, testratio, response)\n",
    "        else:\n",
    "            X_train, X_test, y_train, y_test = sampling.naive_oversample(df, testratio, response)\n",
    "            \n",
    "        if standardize == True:\n",
    "            scaling = MinMaxScaler(feature_range=(-1,1)).fit(X_train)\n",
    "            X_train = scaling.transform(X_train)\n",
    "            X_test = scaling.transform(X_test)\n",
    "            \n",
    "        store = self.build_model(X_train, X_test, y_train, y_test, text, modelname, CV)\n",
    "        # test model with all actual fraud results\n",
    "        store['actual_accuracy'] = evaluate.actual_acc(df, store['model'], response)\n",
    "        return store\n",
    "    \n",
    "    def build_model(self, X_train, X_test, y_train, y_test, text, modelname, CV):\n",
    "        if modelname == 'LogisticRegression':\n",
    "            if CV == True:\n",
    "                param_grid = dict(C=[0.8,1,1.2], max_iter=[300], solver=['liblinear'])\n",
    "                LogRegression = LogisticRegression()\n",
    "                model = GridSearchCV(LogRegression, param_grid, cv=5, scoring='f1', verbose=10)\n",
    "                model.fit(X_train,y_train)\n",
    "                print(\"Best f1 score: \" + str(model.best_score_))\n",
    "                print(\"Best parameters: \" + str(model.best_params_))\n",
    "            else:\n",
    "                model = LogisticRegression(max_iter=300, C=0.8, solver='liblinear')\n",
    "                model.fit(X_train,y_train)\n",
    "        elif modelname == 'XGBoost':\n",
    "            if CV == True:\n",
    "                end_value = math.ceil(math.sqrt(X_train.shape[1]))\n",
    "                start_value = end_value - 2       \n",
    "                # treedepth = list(range(start_value, end_value+1, 2))\n",
    "                param_grid = dict(n_estimators=[100], max_depth=[end_value])\n",
    "                GradientBoost = GradientBoostingClassifier()\n",
    "                model = GridSearchCV(GradientBoost, param_grid, cv=5, scoring='f1', verbose=10)\n",
    "                model.fit(X_train,y_train)\n",
    "                print(\"Best f1 score: \" + str(model.best_score_))\n",
    "                print(\"Best parameters: \" + str(model.best_params_))\n",
    "                \n",
    "                # Testing out xgb.cv (incomplete)\n",
    "                # model = xgb.XGBClassifier(seed=42, nthread=1, max_depth=start_value, n_estimators=100, random_state=42)\n",
    "                # xgb_param = dict(n_estimators=100, max_depth=end_value)\n",
    "                # xgtrain = xgb.DMatrix(X_train, label=y_train)\n",
    "                # model = xgb.cv(params=xgb_param, dtrain=xgtrain, nfold=5, metrics='auc')\n",
    "                # model.fit(X_train, y_train, eval_set=[(X_train, y_train), (X_test, y_test)], verbose=5)\n",
    "                \n",
    "                # USING kfold library to do kfold testing on XGBoost:\n",
    "                # cross_val_score using kfold does not fit the model, so nothing can be predicted\n",
    "                # it's just to see the results but the model has to be fitted later on\n",
    "                # kfold = KFold(n_splits=3, random_state=42)\n",
    "                # print(kfold)\n",
    "                # scores = cross_val_score(model, X_train, y_train, cv=kfold)\n",
    "                # print(\"CV Accuracy: %.2f%% (%.2f%%)\" % (scores.mean()*100, scores.std()*100))\n",
    "            else:\n",
    "                model = xgb.XGBClassifier(seed=42, nthread=1, max_depth=math.ceil(math.sqrt(X_train.shape[1])),\n",
    "                                          n_estimators=100, random_state=42)\n",
    "                model.fit(X_train, y_train, eval_set=[(X_train, y_train), (X_test, y_test)], verbose=5)\n",
    "        elif modelname == 'RandomForest':\n",
    "            if CV == True:\n",
    "                start_value = math.ceil(math.sqrt(X_train.shape[1]))\n",
    "                end_value = start_value + 11         \n",
    "                treedepth = list(range(start_value, end_value, 5))\n",
    "                param_grid = dict(random_state=[42], max_depth=treedepth, n_estimators=[100,150])\n",
    "                RFC = RandomForestClassifier()\n",
    "                model = GridSearchCV(RFC, param_grid, cv=5, scoring='f1', verbose=10)\n",
    "                model.fit(X_train,y_train)\n",
    "                print(\"Best f1 score: \" + str(model.best_score_))\n",
    "                print(\"Best parameters: \" + str(model.best_params_))\n",
    "            else:\n",
    "                treedepth = math.ceil(math.sqrt(X_train.shape[1]))\n",
    "                model = RandomForestClassifier(random_state=42, max_depth=treedepth, n_estimators=150)\n",
    "                model.fit(X_train,y_train)\n",
    "        else:\n",
    "            # Parameters based on gridsearchcv of modelname = logistic regresion\n",
    "            # Leave parameter blank for modelname to run this instance of logistic regression\n",
    "            model = LogisticRegression(C=0.8, max_iter=300, solver='liblinear')\n",
    "            model.fit(X_train,y_train)\n",
    "        \n",
    "        y_predict = model.predict(X_test)\n",
    "        y_predictprob = model.predict_proba(X_test)[:, 1]\n",
    "        results = evaluate.model_results(y_test, y_predict, y_predictprob, text)\n",
    "        store = {\"model\": model, \"X_train\": X_train, \"X_test\": X_test, \"y_train\": y_train, \n",
    "                 \"y_test\": y_test, \"results\": results}\n",
    "        print(\"Model fitting and results are complete!\")\n",
    "        return store\n",
    "    \n",
    "    def standardize(self, df):\n",
    "        # Variables already standardized except for Amount\n",
    "        # columns = df.columns.values.tolist()\n",
    "        # columns.remove(response)\n",
    "        for column in ['Amount']:\n",
    "            df[column] = (df[column] - df[column].mean()) / df[column].std()\n",
    "        return df\n",
    "\n",
    "class sampling:\n",
    "    def __init__(self):\n",
    "        pass\n",
    "    @staticmethod\n",
    "    def naive_oversample(df, testratio, response):\n",
    "        X = df.drop([response], axis=1)\n",
    "        y = df[response]\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=testratio, random_state=41)\n",
    "        ros = RandomOverSampler(random_state=42)\n",
    "        X_train, y_train = ros.fit_resample(X_train, y_train)\n",
    "        # train test split keeps X_test and y_test as pd series, oversampler converts X_train, y_train to numpy\n",
    "        # Convert all to numpy array for XGBoost to not have bugs\n",
    "        X_test = X_test.values\n",
    "        y_test = y_test.values\n",
    "        print(\"Oversampling is complete!\")\n",
    "        return X_train, X_test, y_train, y_test\n",
    "    \n",
    "    @staticmethod\n",
    "    def smote_oversample(df, testratio, response):\n",
    "        X = df.drop([response], axis=1)\n",
    "        y = df[response]\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=testratio, random_state=41)\n",
    "        X_train, y_train = SMOTE().fit_resample(X_train, y_train)\n",
    "        # train test split keeps X_test and y_test as pd series, oversampler converts X_train, y_train to numpy\n",
    "        # Convert all to numpy array for XGBoost to not have bugs\n",
    "        X_test = X_test.values\n",
    "        y_test = y_test.values\n",
    "        print(\"Number of Xs and Ys for SMOTE:\")\n",
    "        print(sorted(Counter(y_train).items()))\n",
    "        print(\"Oversampling is complete!\")\n",
    "        return X_train, X_test, y_train, y_test\n",
    "    \n",
    "    @staticmethod\n",
    "    def adasyn_oversample(df, testratio, response):\n",
    "        X = df.drop([response], axis=1)\n",
    "        y = df[response]\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=testratio, random_state=41)\n",
    "        X_train, y_train = ADASYN().fit_resample(X_train, y_train)\n",
    "        # train test split keeps X_test and y_test as pd series, oversampler converts X_train, y_train to numpy\n",
    "        # Convert all to numpy array for XGBoost to not have bugs\n",
    "        X_test = X_test.values\n",
    "        y_test = y_test.values\n",
    "        print(\"Number of Xs and Ys for ADASYN:\")\n",
    "        print(sorted(Counter(y_train).items()))\n",
    "        print(\"Oversampling is complete!\")\n",
    "        return X_train, X_test, y_train, y_test\n",
    "\n",
    "\n",
    "\n",
    "class evaluate:\n",
    "    def __init__(self):\n",
    "        pass\n",
    "    \n",
    "    @staticmethod\n",
    "    def model_results(y_test, y_predict, y_predictprob, text):\n",
    "        cm = metrics.confusion_matrix(y_test, y_predict)\n",
    "        print(cm)\n",
    "        RFC_CM = pd.DataFrame(cm, ['Actual 0', 'Actual 1'], ['Predict 0', 'Predict 1'])\n",
    "        sns.heatmap(RFC_CM, annot=True, annot_kws={\"size\": 16}, cmap='Greens', linewidths=1, fmt='g')# font size\n",
    "        sns.set(font_scale=1.4)#for label size\n",
    "        plt.title(\"Confusion Matrix for \" + text)\n",
    "\n",
    "        # fix for mpl bug that cuts off top/bottom of seaborn viz\n",
    "        b, t = plt.ylim() \n",
    "        b += 0.5 \n",
    "        t -= 0.5 \n",
    "        plt.ylim(b, t) \n",
    "        plt.show() \n",
    "\n",
    "        accuracy = metrics.accuracy_score(y_test, y_predict)\n",
    "        print('Accuracy: ' + str(accuracy))\n",
    "        sensitivity = cm[1][1] / (cm[1][1] + cm[1][0])\n",
    "        recall = sensitivity\n",
    "        print('Sensitivity: ' + str(sensitivity))\n",
    "        specificity = cm[0][0] / (cm[0][0] + cm[0][1])\n",
    "        print('Specificity: ' + str(specificity))\n",
    "        precision = cm[1][1] / (cm[1][1] + cm[0][1])\n",
    "        print('Precision: ' + str(precision))\n",
    "        f1 = 2 * (recall * precision)/(recall + precision)\n",
    "        print('f1 score: ' + str(f1))\n",
    "        auc, pr_auc = evaluate.ROC(y_test, y_predictprob, text)\n",
    "        results = {\"accuracy\": accuracy, \"sensitivity\": sensitivity, \"specificity\": specificity, \n",
    "                   \"precision\": precision, \"f1\": f1, \"auc\": auc, \"pr_auc\": pr_auc}\n",
    "        print(\"Model classification metrics have finished calculating!\")\n",
    "        print(results)\n",
    "        return results\n",
    "    \n",
    "    @staticmethod\n",
    "    def ROC(y_test, y_predictprob, text):\n",
    "        # IMPORTANT: first argument is true values, second argument is predicted probabilities\n",
    "        auc = metrics.roc_auc_score(y_test, y_predictprob)\n",
    "        # print(\"AUC value is: \" + str(auc))\n",
    "        print(\"AUC value is: \" + str(auc))\n",
    "        fpr, tpr, thresholds = metrics.roc_curve(y_test, y_predictprob)\n",
    "        # print(\"AUC value is also: \" + str(metrics.auc(fpr, tpr)))\n",
    "        # Calculate precision and recall for each threshold\n",
    "        precision, recall, _ = metrics.precision_recall_curve(y_test, y_predictprob)\n",
    "        pr_auc = metrics.auc(recall, precision)\n",
    "        fullgraph = plt.figure(1,figsize=(10,20))\n",
    "        plt.style.use('ggplot')\n",
    "\n",
    "        ROCAUC_plot = fullgraph.add_subplot(211)\n",
    "        ROCAUC_plot.plot(fpr, tpr, color='blue')\n",
    "        ROCAUC_plot.set_title('ROC curve for ' + text)\n",
    "        ROCAUC_plot.set_xlabel('False Positive Rate (1 - Specificity)')\n",
    "        ROCAUC_plot.set_ylabel('True Positive Rate (Sensitivity)')\n",
    "        ROCAUC_plot.set_xlim([0.0, 1.0])\n",
    "        ROCAUC_plot.set_ylim([0.0, 1.0])\n",
    "        ROCAUC_plot.grid(True)\n",
    "        PRAUC_plot = fullgraph.add_subplot(212)\n",
    "        PRAUC_plot.plot(precision, recall, color='purple')\n",
    "        PRAUC_plot.set_title('Precision-Recall curve for ' + text)\n",
    "        PRAUC_plot.set_xlabel('Recall')\n",
    "        PRAUC_plot.set_ylabel('Precision')\n",
    "        PRAUC_plot.set_xlim([0.0, 1.0])\n",
    "        PRAUC_plot.set_ylim([0.0, 1.0])\n",
    "        PRAUC_plot.grid(True)\n",
    "        return auc, pr_auc\n",
    "\n",
    "    @staticmethod\n",
    "    def actual_acc(df, model, response):\n",
    "        allpositive = df[df[response] == 1].copy()\n",
    "        x_positive = allpositive.drop([response], axis=1)\n",
    "        y_positive = allpositive[response]\n",
    "        # Convert to numpy array due to XGBoost model.predict not working well for pandas\n",
    "        x_positive = x_positive.values\n",
    "        y_positive = y_positive.values\n",
    "        y_pospredict = model.predict(x_positive)\n",
    "        accuracy_positive = metrics.accuracy_score(y_positive, y_pospredict)\n",
    "        print(\"Accuracy with all fraud results is \" + str(accuracy_positive * 100) + \"%\")\n",
    "        return accuracy_positive"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initialize pipeline object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "modelpipeline = modelpipeline()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test that oversampling works"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of Xs and Ys for SMOTE:\n",
      "[(0, 227447), (1, 227447)]\n",
      "Oversampling is complete!\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = sampling.smote_oversample(df,0.2,'Class')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "227447"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "94"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import statsmodels.discrete.discrete_model as sm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Testing some functions and tricks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "V1\n",
      "V2\n",
      "V3\n",
      "V4\n",
      "V5\n",
      "V6\n",
      "V7\n",
      "V8\n",
      "V9\n",
      "V10\n",
      "V11\n",
      "V12\n",
      "V13\n",
      "V14\n",
      "V15\n",
      "V16\n",
      "V17\n",
      "V18\n",
      "V19\n",
      "V20\n",
      "V21\n",
      "V22\n",
      "V23\n",
      "V24\n",
      "V25\n",
      "V26\n",
      "V27\n",
      "V28\n",
      "Amount\n",
      "Class\n"
     ]
    }
   ],
   "source": [
    "for colname in df.columns.values:\n",
    "    print(colname)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(454894, 29)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-3.60900513e-01,  1.28452281e+00, -1.01771767e+00, ...,\n",
       "        -1.32423897e-01,  1.06391985e-01,  9.95000000e+00],\n",
       "       [-5.41761567e-01,  1.18792390e+00,  9.43541216e-01, ...,\n",
       "         3.44136513e-01,  1.51161024e-01,  8.90000000e-01],\n",
       "       [-2.58931571e-01,  4.35045463e-01,  1.48869152e+00, ...,\n",
       "        -1.52822890e-01, -9.01466370e-02,  6.90000000e+00],\n",
       "       ...,\n",
       "       [-1.08943916e+00,  1.44952753e+00, -1.61685654e+00, ...,\n",
       "         4.99160527e-01,  2.89658112e-01,  7.74400674e+01],\n",
       "       [-1.75122393e+01,  1.01499441e+01, -2.42394972e+01, ...,\n",
       "         2.09205512e+00, -1.42571245e+00,  1.00000000e+00],\n",
       "       [-4.16356602e+00,  2.66418080e+00, -3.37813493e+00, ...,\n",
       "        -1.10787844e+00, -1.07434974e-01,  2.45035838e+02]])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[:,0:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(454894, 29)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.00000000e+00, -3.60900513e-01,  1.28452281e+00, ...,\n",
       "        -1.32423897e-01,  1.06391985e-01,  9.95000000e+00],\n",
       "       [ 0.00000000e+00, -5.41761567e-01,  1.18792390e+00, ...,\n",
       "         3.44136513e-01,  1.51161024e-01,  8.90000000e-01],\n",
       "       [ 0.00000000e+00, -2.58931571e-01,  4.35045463e-01, ...,\n",
       "        -1.52822890e-01, -9.01466370e-02,  6.90000000e+00],\n",
       "       ...,\n",
       "       [ 0.00000000e+00, -1.08943916e+00,  1.44952753e+00, ...,\n",
       "         4.99160527e-01,  2.89658112e-01,  7.74400674e+01],\n",
       "       [ 0.00000000e+00, -1.75122393e+01,  1.01499441e+01, ...,\n",
       "         2.09205512e+00, -1.42571245e+00,  1.00000000e+00],\n",
       "       [ 0.00000000e+00, -4.16356602e+00,  2.66418080e+00, ...,\n",
       "        -1.10787844e+00, -1.07434974e-01,  2.45035838e+02]])"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "full_x = np.empty((454894,1))\n",
    "np.concatenate((full_x, X_train), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('c', 1000)\n",
      "c\n",
      "1000\n"
     ]
    }
   ],
   "source": [
    "# Finding element of dictionary with max value\n",
    "import operator\n",
    "stats = {\"a\":1, \"b\":2, \"c\":1000}\n",
    "print(max(stats.items(), key=operator.itemgetter(1)))\n",
    "print(max(stats.items(), key=operator.itemgetter(1))[0])\n",
    "print(max(stats.items(), key=operator.itemgetter(1))[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.zeros((1,1)).any()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array(False).any()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Actual implementation of forward selection"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Total features to select = total number in colnames array = k\n",
    "### In each iteration, the current set of n features is concatenated with a new feature not inside current set\n",
    "### It is then sent for training with the logistic regression\n",
    "### The model performance for each feature + current features is evaluated by its highest p value (worst feature)\n",
    "### All highest p values of all feature addition to n features (k-n iterations) are put into a dictionary\n",
    "### Next, the lowest p value out of all the iterations (for n features + 1) is chosen for evaluation\n",
    "### Set significance level, which is compared to the lowest p value of the best model in the current training iteration\n",
    "### If best model in current training iteration of n vars has any vars with p value > sig level, then the model training stops\n",
    "### Because all the different models are worse or equally bad as the current best model, we can terminate selection process\n",
    "### If not, repeat this iteration with now n+1 features and k-n-1 iterations\n",
    "### Repeat until there are no more better models or all features are selected"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount']\n",
      "i is 0\n",
      "j is 0\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "()\n",
      "False\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.564092\n",
      "         Iterations 7\n",
      "First model trained using feature number 0 with p value of 0\n",
      "Full X:\n",
      "(454894, 1)\n",
      "i is 1\n",
      "j is 1\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 1)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.510754\n",
      "         Iterations 7\n",
      "Features all have p value of 0, using feature number 1\n",
      "i is 2\n",
      "j is 2\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 2)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.448630\n",
      "         Iterations 8\n",
      "Features all have p value of 0, using feature number 2\n",
      "i is 3\n",
      "j is 3\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 3)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.422452\n",
      "         Iterations 8\n",
      "Features all have p value of 0, using feature number 3\n",
      "i is 4\n",
      "j is 4\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 4)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420759\n",
      "         Iterations 9\n",
      "Features all have p value of 0, using feature number 4\n",
      "i is 5\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.419426\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 6\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.415607\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 7\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420709\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 8\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.418189\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.404809\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 10\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.408897\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 11\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.403220\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420265\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 13\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.391059\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420753\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 15\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.414738\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 16\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.416288\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420174\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420758\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 19\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420743\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420062\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420759\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 22\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420094\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420617\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420570\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420715\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 26\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.419883\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420260\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 28\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.364210\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 28\n",
      "i is 6\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.363716\n",
      "         Iterations 10\n",
      "i is 6\n",
      "j is 6\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.357805\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 6\n",
      "i is 7\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.356400\n",
      "         Iterations 10\n",
      "i is 7\n",
      "j is 7\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.353298\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 7\n",
      "i is 8\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 8)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.351768\n",
      "         Iterations 10\n",
      "i is 8\n",
      "j is 8\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 8)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.348619\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 8\n",
      "i is 9\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 9)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.346589\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 5\n",
      "i is 10\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 10)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.286862\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 9\n",
      "i is 11\n",
      "j is 10\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 11)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.279469\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 10\n",
      "i is 12\n",
      "j is 11\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 12)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.263986\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 11\n",
      "i is 13\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 13)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.263935\n",
      "         Iterations 11\n",
      "i is 13\n",
      "j is 13\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 13)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.250935\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 13\n",
      "i is 14\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.250933\n",
      "         Iterations 11\n",
      "i is 14\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.250900\n",
      "         Iterations 11\n",
      "i is 14\n",
      "j is 15\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.248668\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 15\n",
      "i is 15\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 15)\n",
      "True\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.248667\n",
      "         Iterations 11\n",
      "i is 15\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 15)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.248635\n",
      "         Iterations 11\n",
      "i is 15\n",
      "j is 16\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 15)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.236450\n",
      "         Iterations 13\n",
      "Features all have p value of 0, using feature number 16\n",
      "i is 16\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 16)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.236448\n",
      "         Iterations 13\n",
      "i is 16\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 16)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.236423\n",
      "         Iterations 13\n",
      "i is 16\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 16)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.234328\n",
      "         Iterations 15\n",
      "Features all have p value of 0, using feature number 17\n",
      "i is 17\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.234327\n",
      "         Iterations 15\n",
      "i is 17\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.234316\n",
      "         Iterations 15\n",
      "i is 17\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.234140\n",
      "         Iterations 15\n",
      "i is 17\n",
      "j is 19\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212927\n",
      "         Iterations 15\n",
      "Features all have p value of 0, using feature number 19\n",
      "i is 18\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212816\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212761\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212413\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.200371\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212416\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 22\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.201928\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212788\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211543\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212592\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 26\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211394\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211375\n",
      "         Iterations 15\n",
      "Getting new model with min p-values with 18 variables.\n",
      "New model trained using feature number 26 with lowest p values of 4.364524237978316e-293\n",
      "i is 19\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211284\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211267\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210911\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.198853\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210870\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 22\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.199119\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211267\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210111\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211128\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210396\n",
      "         Iterations 15\n",
      "Getting new model with min p-values with 19 variables.\n",
      "New model trained using feature number 22 with lowest p values of 2.157027477523322e-264\n",
      "i is 20\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.199070\n",
      "         Iterations 15\n",
      "i is 20\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.199116\n",
      "         Iterations 15\n",
      "i is 20\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.198079\n",
      "         Iterations 15\n",
      "i is 20\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.185037\n",
      "         Iterations 15\n",
      "Features all have p value of 0, using feature number 20\n",
      "i is 21\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.184963\n",
      "         Iterations 15\n",
      "i is 21\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.185037\n",
      "         Iterations 15\n",
      "i is 21\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.184317\n",
      "         Iterations 16\n",
      "i is 21\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.182719\n",
      "         Iterations 15\n",
      "Features all have p value of 0, using feature number 21\n",
      "i is 22\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.182600\n",
      "         Iterations 15\n",
      "i is 22\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.182717\n",
      "         Iterations 15\n",
      "i is 22\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.181863\n",
      "         Iterations 15\n",
      "i is 22\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.182716\n",
      "         Iterations 15\n",
      "i is 22\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.178477\n",
      "         Iterations 16\n",
      "i is 22\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.182645\n",
      "         Iterations 15\n",
      "i is 22\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.180289\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 22 variables.\n",
      "New model trained using feature number 24 with lowest p values of 1.0447168325500864e-296\n",
      "i is 23\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.178318\n",
      "         Iterations 16\n",
      "i is 23\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.178475\n",
      "         Iterations 16\n",
      "i is 23\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.177006\n",
      "         Iterations 15\n",
      "i is 23\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.178477\n",
      "         Iterations 16\n",
      "i is 23\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.178319\n",
      "         Iterations 16\n",
      "i is 23\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.176160\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 23 variables.\n",
      "New model trained using feature number 27 with lowest p values of 8.029805300654864e-235\n",
      "i is 24\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.175961\n",
      "         Iterations 16\n",
      "i is 24\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.176160\n",
      "         Iterations 16\n",
      "i is 24\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174591\n",
      "         Iterations 16\n",
      "i is 24\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.176160\n",
      "         Iterations 16\n",
      "i is 24\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.175997\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 24 variables.\n",
      "New model trained using feature number 18 with lowest p values of 2.799540855543493e-144\n",
      "i is 25\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 25)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174323\n",
      "         Iterations 16\n",
      "i is 25\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 25)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174574\n",
      "         Iterations 16\n",
      "i is 25\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 25)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174591\n",
      "         Iterations 16\n",
      "i is 25\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 25)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174492\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 25 variables.\n",
      "New model trained using feature number 12 with lowest p values of 6.744714932826996e-55\n",
      "i is 26\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 26)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174313\n",
      "         Iterations 16\n",
      "i is 26\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 26)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174323\n",
      "         Iterations 16\n",
      "i is 26\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 26)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174227\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 26 variables.\n",
      "New model trained using feature number 25 with lowest p values of 1.561502837087903e-20\n",
      "i is 27\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 27)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174217\n",
      "         Iterations 16\n",
      "i is 27\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 27)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174227\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 27 variables.\n",
      "New model trained using feature number 14 with lowest p values of 0.0028516219719427064\n",
      "i is 28\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 28)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174217\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 28 variables.\n",
      "TERMINATING AS best model trained using feature number 23 with high p value of 0.7040606741302529 above significance level: 0.05\n",
      "['V1', 'V2', 'V3', 'V4', 'V5', 'Amount', 'V7', 'V8', 'V9', 'V6', 'V10', 'V11', 'V12', 'V14', 'V16', 'V17', 'V18', 'V20', 'V27', 'V23', 'V21', 'V22', 'V25', 'V28', 'V19', 'V13', 'V26', 'V15']\n"
     ]
    }
   ],
   "source": [
    "import operator\n",
    "import statsmodels.discrete.discrete_model as sm\n",
    "\n",
    "# Column names have Time (irrelevant feature) and Class (response) removed\n",
    "colnames = list(df.columns.values)\n",
    "colnames.remove('Class')\n",
    "print(colnames)\n",
    "\n",
    "# Total features to select = total number in colnames array = k\n",
    "# In each iteration, the current set of n features is concatenated with a new feature not inside current set\n",
    "# It is then sent for training with the logistic regression\n",
    "# The model performance for each feature + current features is evaluated by its highest p value (worst feature)\n",
    "# All highest p values of all feature addition to n features (k-n iterations) are put into a dictionary\n",
    "# Next, the lowest p value out of all the iterations (for n features + 1) is chosen for evaluation\n",
    "# Set significance level, which is compared to the lowest p value of the best model in the current training iteration\n",
    "# If best model in current training iteration of n vars has any vars with p value > sig level, then the model training stops\n",
    "# Because all the different models are worse or equally bad as the current best model, we can terminate selection process\n",
    "# If not, repeat this iteration with now n+1 features and k-n-1 iterations\n",
    "# Repeat until there are no more better models or all features are selected\n",
    "sig_level = 0.05\n",
    "\n",
    "maxcolsnum = X_train.shape[1]\n",
    "full_x = np.array(False)\n",
    "allowed_nums = list(range(maxcolsnum))\n",
    "actual_nums = []\n",
    "actual_vars = []\n",
    "terminate_early = False\n",
    "y = y_train\n",
    "for i in range(maxcolsnum):\n",
    "    # Reset boolean and pval_list\n",
    "    terminate_early = False\n",
    "    pval_list = {}\n",
    "    for j in range(maxcolsnum):\n",
    "        if j in allowed_nums:\n",
    "            # Need to reshape to single column instead of a long array for concating properly\n",
    "            jth_x = X_train[:,j].reshape(-1,1)\n",
    "            print(\"i is \" + str(i))\n",
    "            print(\"j is \" + str(j))\n",
    "            print(\"Jth_x:\")\n",
    "            print(jth_x.shape)\n",
    "            print(\"Full_x:\")\n",
    "            print(full_x.shape)\n",
    "            print(full_x.any())\n",
    "            if full_x.any():\n",
    "                iter_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "            else:\n",
    "                iter_x = jth_x\n",
    "            regressor_OLS = sm.Logit(y_train, iter_x).fit()\n",
    "            # print(\"p value for column number: \" + str(i))\n",
    "            # print(regressor_OLS.pvalues[0])\n",
    "            pval_list[j] = max(regressor_OLS.pvalues)\n",
    "            # Special condition where all the features have p values of 0, directly use these variables for training\n",
    "            if max(regressor_OLS.pvalues) == 0:\n",
    "                if full_x.any():\n",
    "                    full_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "                    allowed_nums.remove(j)\n",
    "                    actual_nums.append(j)\n",
    "                    print(\"Features all have p value of 0, using feature number \" + str(j))\n",
    "                else:\n",
    "                    full_x = jth_x\n",
    "                    allowed_nums.remove(j)\n",
    "                    actual_nums.append(j)\n",
    "                    print(\"First model trained using feature number \" + str(j) + \" with p value of 0\")\n",
    "                    print(\"Full X:\")\n",
    "                    print(full_x.shape)\n",
    "                terminate_early = True\n",
    "                break\n",
    "        else:\n",
    "            continue\n",
    "    if i > 0 and terminate_early == False:\n",
    "        print(\"Getting new model with min p-values with \" + str(len(actual_nums)) + \" variables.\")\n",
    "        max_pval_col = min(pval_list.items(), key=operator.itemgetter(1))[0]\n",
    "        max_pval = pval_list[max_pval_col]\n",
    "        # Need to reshape to single column instead of a long array for concating properly\n",
    "        jth_x = X_train[:,max_pval_col].reshape(-1,1)\n",
    "        if max_pval < sig_level:\n",
    "            if full_x.any():\n",
    "                full_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "                allowed_nums.remove(max_pval_col)\n",
    "                actual_nums.append(max_pval_col)\n",
    "                print(\"New model trained using feature number \" + str(max_pval_col) + \" with lowest p values of \" + str(max_pval))\n",
    "            else:\n",
    "                full_x = jth_x\n",
    "                allowed_nums.remove(max_pval_col)\n",
    "                actual_nums.append(max_pval_col)\n",
    "                print(\"First model trained using feature number \" + str(max_pval_col) + \" with lowest p values of \" + str(max_pval))\n",
    "        else:\n",
    "            print(\"TERMINATING AS best model trained using feature number \" + str(max_pval_col) + \" with high p value of \" + str(max_pval) + \" above significance level: \" + str(sig_level))\n",
    "            break\n",
    "                \n",
    "\n",
    "\n",
    "for k in actual_nums:\n",
    "    actual_vars.append(colnames[k])\n",
    "print(actual_vars)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Basic forward selection with a dataframe but not in a function"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Some minor improvements to previous code above but the reusability of the code is low as it is not made into a function yet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount']\n",
      "i is 0\n",
      "j is 0\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "()\n",
      "False\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.564092\n",
      "         Iterations 7\n",
      "First model trained using feature number 0 with p value of 0\n",
      "Full X:\n",
      "(454894, 1)\n",
      "i is 1\n",
      "j is 1\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 1)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.510754\n",
      "         Iterations 7\n",
      "Features all have p value of 0, using feature number 1\n",
      "i is 2\n",
      "j is 2\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 2)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.448630\n",
      "         Iterations 8\n",
      "Features all have p value of 0, using feature number 2\n",
      "i is 3\n",
      "j is 3\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 3)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.422452\n",
      "         Iterations 8\n",
      "Features all have p value of 0, using feature number 3\n",
      "i is 4\n",
      "j is 4\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 4)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420759\n",
      "         Iterations 9\n",
      "Features all have p value of 0, using feature number 4\n",
      "i is 5\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.419426\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 6\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.415607\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 7\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420709\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 8\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.418189\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.404809\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 10\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.408897\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 11\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.403220\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420265\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 13\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.391059\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420753\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 15\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.414738\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 16\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.416288\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420174\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420758\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 19\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420743\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420062\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420759\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 22\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420094\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420617\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420570\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420715\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 26\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.419883\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420260\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 28\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.364210\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 28\n",
      "i is 6\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.363716\n",
      "         Iterations 10\n",
      "i is 6\n",
      "j is 6\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.357805\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 6\n",
      "i is 7\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.356400\n",
      "         Iterations 10\n",
      "i is 7\n",
      "j is 7\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.353298\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 7\n",
      "i is 8\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 8)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.351768\n",
      "         Iterations 10\n",
      "i is 8\n",
      "j is 8\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 8)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.348619\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 8\n",
      "i is 9\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 9)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.346589\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 5\n",
      "i is 10\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 10)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.286862\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 9\n",
      "i is 11\n",
      "j is 10\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 11)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.279469\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 10\n",
      "i is 12\n",
      "j is 11\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 12)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.263986\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 11\n",
      "i is 13\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 13)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.263935\n",
      "         Iterations 11\n",
      "i is 13\n",
      "j is 13\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 13)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.250935\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 13\n",
      "i is 14\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.250933\n",
      "         Iterations 11\n",
      "i is 14\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.250900\n",
      "         Iterations 11\n",
      "i is 14\n",
      "j is 15\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.248668\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 15\n",
      "i is 15\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 15)\n",
      "True\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.248667\n",
      "         Iterations 11\n",
      "i is 15\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 15)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.248635\n",
      "         Iterations 11\n",
      "i is 15\n",
      "j is 16\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 15)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.236450\n",
      "         Iterations 13\n",
      "Features all have p value of 0, using feature number 16\n",
      "i is 16\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 16)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.236448\n",
      "         Iterations 13\n",
      "i is 16\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 16)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.236423\n",
      "         Iterations 13\n",
      "i is 16\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 16)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.234328\n",
      "         Iterations 15\n",
      "Features all have p value of 0, using feature number 17\n",
      "i is 17\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.234327\n",
      "         Iterations 15\n",
      "i is 17\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.234316\n",
      "         Iterations 15\n",
      "i is 17\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.234140\n",
      "         Iterations 15\n",
      "i is 17\n",
      "j is 19\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212927\n",
      "         Iterations 15\n",
      "Features all have p value of 0, using feature number 19\n",
      "i is 18\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212816\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212761\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212413\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.200371\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212416\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 22\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.201928\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212788\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211543\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212592\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 26\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211394\n",
      "         Iterations 15\n",
      "i is 18\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211375\n",
      "         Iterations 15\n",
      "Getting new model with min p-values with 18 variables.\n",
      "New model trained using feature number 26 with lowest p values of 4.364524237978316e-293\n",
      "i is 19\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211284\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211267\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210911\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.198853\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210870\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 22\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.199119\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211267\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210111\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.211128\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210396\n",
      "         Iterations 15\n",
      "Getting new model with min p-values with 19 variables.\n",
      "New model trained using feature number 22 with lowest p values of 2.157027477523322e-264\n",
      "i is 20\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.199070\n",
      "         Iterations 15\n",
      "i is 20\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.199116\n",
      "         Iterations 15\n",
      "i is 20\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.198079\n",
      "         Iterations 15\n",
      "i is 20\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.185037\n",
      "         Iterations 15\n",
      "Features all have p value of 0, using feature number 20\n",
      "i is 21\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.184963\n",
      "         Iterations 15\n",
      "i is 21\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.185037\n",
      "         Iterations 15\n",
      "i is 21\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.184317\n",
      "         Iterations 16\n",
      "i is 21\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.182719\n",
      "         Iterations 15\n",
      "Features all have p value of 0, using feature number 21\n",
      "i is 22\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.182600\n",
      "         Iterations 15\n",
      "i is 22\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.182717\n",
      "         Iterations 15\n",
      "i is 22\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.181863\n",
      "         Iterations 15\n",
      "i is 22\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.182716\n",
      "         Iterations 15\n",
      "i is 22\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.178477\n",
      "         Iterations 16\n",
      "i is 22\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.182645\n",
      "         Iterations 15\n",
      "i is 22\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.180289\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 22 variables.\n",
      "New model trained using feature number 24 with lowest p values of 1.0447168325500864e-296\n",
      "i is 23\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.178318\n",
      "         Iterations 16\n",
      "i is 23\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.178475\n",
      "         Iterations 16\n",
      "i is 23\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.177006\n",
      "         Iterations 15\n",
      "i is 23\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.178477\n",
      "         Iterations 16\n",
      "i is 23\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.178319\n",
      "         Iterations 16\n",
      "i is 23\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.176160\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 23 variables.\n",
      "New model trained using feature number 27 with lowest p values of 8.029805300654864e-235\n",
      "i is 24\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.175961\n",
      "         Iterations 16\n",
      "i is 24\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.176160\n",
      "         Iterations 16\n",
      "i is 24\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174591\n",
      "         Iterations 16\n",
      "i is 24\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.176160\n",
      "         Iterations 16\n",
      "i is 24\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.175997\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 24 variables.\n",
      "New model trained using feature number 18 with lowest p values of 2.799540855543493e-144\n",
      "i is 25\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 25)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174323\n",
      "         Iterations 16\n",
      "i is 25\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 25)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174574\n",
      "         Iterations 16\n",
      "i is 25\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 25)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174591\n",
      "         Iterations 16\n",
      "i is 25\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 25)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174492\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 25 variables.\n",
      "New model trained using feature number 12 with lowest p values of 6.744714932826996e-55\n",
      "i is 26\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 26)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174313\n",
      "         Iterations 16\n",
      "i is 26\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 26)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174323\n",
      "         Iterations 16\n",
      "i is 26\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 26)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174227\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 26 variables.\n",
      "New model trained using feature number 25 with lowest p values of 1.561502837087903e-20\n",
      "i is 27\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 27)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174217\n",
      "         Iterations 16\n",
      "i is 27\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 27)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174227\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 27 variables.\n",
      "New model trained using feature number 14 with lowest p values of 0.0028516219719427064\n",
      "i is 28\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454894, 1)\n",
      "Full_x:\n",
      "(454894, 28)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174217\n",
      "         Iterations 16\n",
      "Getting new model with min p-values with 28 variables.\n",
      "TERMINATING AS best model trained using feature number 23 with high p value of 0.7040606741302529 above significance level: 0.05\n",
      "Final variables selected:\n",
      "['V1', 'V2', 'V3', 'V4', 'V5', 'Amount', 'V7', 'V8', 'V9', 'V6', 'V10', 'V11', 'V12', 'V14', 'V16', 'V17', 'V18', 'V20', 'V27', 'V23', 'V21', 'V22', 'V25', 'V28', 'V19', 'V13', 'V26', 'V15']\n"
     ]
    }
   ],
   "source": [
    "import operator\n",
    "import statsmodels.discrete.discrete_model as sm\n",
    "\n",
    "# Column names have Time (irrelevant feature) and Class (response) removed\n",
    "colnames = list(df.columns.values)\n",
    "colnames.remove('Class')\n",
    "print(colnames)\n",
    "\n",
    "# Total features to select = total number in colnames array = k\n",
    "# In each iteration, the current set of n features is concatenated with a new feature not inside current set\n",
    "# It is then sent for training with the logistic regression\n",
    "# The model performance for each feature + current features is evaluated by its highest p value (worst feature)\n",
    "# All highest p values of all feature addition to n features (k-n iterations) are put into a dictionary\n",
    "# Next, the lowest p value out of all the iterations (for n features + 1) is chosen for evaluation\n",
    "# Set significance level, which is compared to the lowest p value of the best model in the current training iteration\n",
    "# If best model in current training iteration of n vars has any vars with p value > sig level, then the model training stops\n",
    "# Because all the different models are worse or equally bad as the current best model, we can terminate selection process\n",
    "# If not, repeat this iteration with now n+1 features and k-n-1 iterations\n",
    "# Repeat until there are no more better models or all features are selected\n",
    "sig_level = 0.05\n",
    "\n",
    "maxcolsnum = X_train.shape[1]\n",
    "full_x = np.array(False)\n",
    "allowed_nums = {}\n",
    "for i in range(maxcolsnum):\n",
    "    allowed_nums[i] = True\n",
    "actual_nums = []\n",
    "actual_vars = []\n",
    "terminate_early = False\n",
    "y = y_train\n",
    "for i in range(maxcolsnum):\n",
    "    # Reset boolean and pval_list\n",
    "    terminate_early = False\n",
    "    pval_list = {}\n",
    "    for j in range(maxcolsnum):\n",
    "        if allowed_nums[j] == True:\n",
    "            # Need to reshape to single column instead of a long array for concating properly\n",
    "            jth_x = X_train[:,j].reshape(-1,1)\n",
    "            print(\"i is \" + str(i))\n",
    "            print(\"j is \" + str(j))\n",
    "            print(\"Jth_x:\")\n",
    "            print(jth_x.shape)\n",
    "            print(\"Full_x:\")\n",
    "            print(full_x.shape)\n",
    "            print(full_x.any())\n",
    "            if full_x.any():\n",
    "                iter_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "            else:\n",
    "                iter_x = jth_x\n",
    "            regressor_OLS = sm.Logit(y_train, iter_x).fit()\n",
    "            # print(\"p value for column number: \" + str(i))\n",
    "            # print(regressor_OLS.pvalues[0])\n",
    "            pval_list[j] = max(regressor_OLS.pvalues)\n",
    "            # Special condition where all the features have p values of 0, directly use these variables for training\n",
    "            if max(regressor_OLS.pvalues) == 0:\n",
    "                if full_x.any():\n",
    "                    full_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "                    allowed_nums[j] = False\n",
    "                    actual_nums.append(j)\n",
    "                    print(\"Features all have p value of 0, using feature number \" + str(j))\n",
    "                else:\n",
    "                    full_x = jth_x\n",
    "                    allowed_nums[j] = False\n",
    "                    actual_nums.append(j)\n",
    "                    print(\"First model trained using feature number \" + str(j) + \" with p value of 0\")\n",
    "                    print(\"Full X:\")\n",
    "                    print(full_x.shape)\n",
    "                terminate_early = True\n",
    "                break\n",
    "        else:\n",
    "            continue\n",
    "    if i > 0 and terminate_early == False:\n",
    "        print(\"Getting new model with min p-values with \" + str(len(actual_nums)) + \" variables.\")\n",
    "        max_pval_col = min(pval_list.items(), key=operator.itemgetter(1))[0]\n",
    "        max_pval = pval_list[max_pval_col]\n",
    "        # Need to reshape to single column instead of a long array for concating properly\n",
    "        jth_x = X_train[:,max_pval_col].reshape(-1,1)\n",
    "        if max_pval < sig_level:\n",
    "            if full_x.any():\n",
    "                full_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "                allowed_nums[max_pval_col] = False\n",
    "                actual_nums.append(max_pval_col)\n",
    "                print(\"New model trained using feature number \" + str(max_pval_col) + \" with lowest p values of \" + str(max_pval))\n",
    "            else:\n",
    "                full_x = jth_x\n",
    "                allowed_nums[max_pval_col] = False\n",
    "                actual_nums.append(max_pval_col)\n",
    "                print(\"First model trained using feature number \" + str(max_pval_col) + \" with lowest p values of \" + str(max_pval))\n",
    "        else:\n",
    "            print(\"TERMINATING AS best model trained using feature number \" + str(max_pval_col) + \" with high p value of \" + str(max_pval) + \" above significance level: \" + str(sig_level))\n",
    "            break\n",
    "                \n",
    "\n",
    "for k in actual_nums:\n",
    "    actual_vars.append(colnames[k])\n",
    "print('Final variables selected:')\n",
    "print(actual_vars)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Forward selection function\n",
    "## Accepts a dataframe and does train-test split, along with oversampling\n",
    "### Code is highly reusable for different kinds of classification problems with imbalanced dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import operator\n",
    "from collections import Counter\n",
    "from imblearn.over_sampling import RandomOverSampler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from imblearn.over_sampling import ADASYN\n",
    "import statsmodels.discrete.discrete_model as sm\n",
    "\n",
    "def forwardselection(df,testratio,response,removelist,sampling):\n",
    "    if isinstance(removelist, str) == True:\n",
    "        temp_str = removelist\n",
    "        removelist = []\n",
    "        removelist.append(temp_str)\n",
    "    X = df.drop(removelist, axis=1)\n",
    "    y = df[response]\n",
    "    # Get list of column names\n",
    "    colnames = list(X.columns.values)\n",
    "    print(colnames)\n",
    "    \n",
    "    # Start of train-test split and oversampling (if relevant)\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=testratio, random_state=42)\n",
    "    if sampling.lower() == 'smote':\n",
    "        print(\"SMOTE Oversampling selected..\")\n",
    "        X_train, y_train = SMOTE().fit_resample(X_train, y_train)\n",
    "        # train test split keeps X_test and y_test as pd series, oversampler converts X_train, y_train to numpy\n",
    "        # Convert all to numpy array for XGBoost to not have bugs\n",
    "        X_test = X_test.values\n",
    "        y_test = y_test.values\n",
    "        print(\"Number of Xs and Ys for: \" + str(sampling.upper()))\n",
    "        print(sorted(Counter(y_train).items()))\n",
    "        print(\"Oversampling is complete!\")\n",
    "    elif sampling.lower() == 'naive':\n",
    "        print(\"Naive Oversampling selected..\")\n",
    "        ros = RandomOverSampler(random_state=42)\n",
    "        X_train, y_train = ros.fit_resample(X_train, y_train)\n",
    "        # train test split keeps X_test and y_test as pd series, oversampler converts X_train, y_train to numpy\n",
    "        # Convert all to numpy array for XGBoost to not have bugs\n",
    "        X_test = X_test.values\n",
    "        y_test = y_test.values\n",
    "        print(\"Number of Xs and Ys for: \" + str(sampling.upper()))\n",
    "        print(sorted(Counter(y_train).items()))\n",
    "        print(\"Oversampling is complete!\")\n",
    "\n",
    "\n",
    "    # Total features to select = k\n",
    "    # In each iteration, the current set of n features is concatenated with a new feature not inside current set\n",
    "    # It is then sent for training with the logistic regression\n",
    "    # The model performance for each feature + current features is evaluated by its highest p value (worst feature)\n",
    "    # All highest p values of all feature addition to n features (k-n iterations) are put into a dictionary\n",
    "    # Next, the lowest p value out of all the iterations (for n features + 1) is chosen for evaluation\n",
    "    # Set significance level, which is compared to the lowest p value of the best model in the current training iteration\n",
    "    # If best model in current training iteration of n vars has any vars with p value > sig level, then the model training stops\n",
    "    # Because all the different models are worse or equally bad as the current best model, we can terminate selection process\n",
    "    # If not, repeat this iteration with now n+1 features and k-n-1 iterations\n",
    "    sig_level = 0.05\n",
    "\n",
    "    maxcolsnum = X_train.shape[1]\n",
    "    full_x = np.array(False)\n",
    "    allowed_nums = {}\n",
    "    for i in range(maxcolsnum):\n",
    "        allowed_nums[i] = True\n",
    "    actual_nums = []\n",
    "    actual_vars = []\n",
    "    terminate_early = False\n",
    "    y = y_train\n",
    "    for i in range(maxcolsnum):\n",
    "        # Reset boolean and pval_list\n",
    "        terminate_early = False\n",
    "        pval_list = {}\n",
    "        for j in range(maxcolsnum):\n",
    "            if allowed_nums[j] == True:\n",
    "                # Need to reshape to single column instead of a long array for concating properly\n",
    "                jth_x = X_train[:,j].reshape(-1,1)\n",
    "                print(\"i is \" + str(i))\n",
    "                print(\"j is \" + str(j))\n",
    "                print(\"Jth_x:\")\n",
    "                print(jth_x.shape)\n",
    "                print(\"Full_x:\")\n",
    "                print(full_x.shape)\n",
    "                print(full_x.any())\n",
    "                if full_x.any():\n",
    "                    iter_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "                else:\n",
    "                    iter_x = jth_x\n",
    "                regressor_OLS = sm.Logit(y_train, iter_x).fit()\n",
    "                # print(\"p value for column number: \" + str(i))\n",
    "                # print(regressor_OLS.pvalues[0])\n",
    "                pval_list[j] = max(regressor_OLS.pvalues)\n",
    "                # Special condition where all the features have p values of 0, directly use these variables for training\n",
    "                if max(regressor_OLS.pvalues) == 0:\n",
    "                    if full_x.any():\n",
    "                        full_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "                        allowed_nums[j] = False\n",
    "                        actual_nums.append(j)\n",
    "                        print(\"Features all have p value of 0, using feature number \" + str(j))\n",
    "                    else:\n",
    "                        full_x = jth_x\n",
    "                        allowed_nums[j] = False\n",
    "                        actual_nums.append(j)\n",
    "                        print(\"First model trained using feature number \" + str(j) + \" with p value of 0\")\n",
    "                        print(\"Full X:\")\n",
    "                        print(full_x.shape)\n",
    "                    terminate_early = True\n",
    "                    break\n",
    "            else:\n",
    "                continue\n",
    "        if i > 0 and terminate_early == False:\n",
    "            print(\"Getting new model with min p-values with \" + str(len(actual_nums)) + \" variables.\")\n",
    "            max_pval_col = min(pval_list.items(), key=operator.itemgetter(1))[0]\n",
    "            max_pval = pval_list[max_pval_col]\n",
    "            # Need to reshape to single column instead of a long array for concating properly\n",
    "            jth_x = X_train[:,max_pval_col].reshape(-1,1)\n",
    "            if max_pval < sig_level:\n",
    "                if full_x.any():\n",
    "                    full_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "                    allowed_nums[max_pval_col] = False\n",
    "                    actual_nums.append(max_pval_col)\n",
    "                    print(\"New model trained using feature number \" + str(max_pval_col) + \" with lowest p values of \" + str(max_pval))\n",
    "                else:\n",
    "                    full_x = jth_x\n",
    "                    allowed_nums[max_pval_col] = False\n",
    "                    actual_nums.append(max_pval_col)\n",
    "                    print(\"First model trained using feature number \" + str(max_pval_col) + \" with lowest p values of \" + str(max_pval))\n",
    "            else:\n",
    "                print(\"TERMINATING AS best model trained using feature number \" + str(max_pval_col) + \" with high p value of \" + str(max_pval) + \" above significance level: \" + str(sig_level))\n",
    "                break\n",
    "\n",
    "\n",
    "    for k in actual_nums:\n",
    "        actual_vars.append(colnames[k])\n",
    "    print('Final variables selected:')\n",
    "    print(actual_vars)\n",
    "    \n",
    "    return actual_vars"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>Time</th>\n",
       "      <th>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>V9</th>\n",
       "      <th>...</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>186521</td>\n",
       "      <td>127145.0</td>\n",
       "      <td>0.077780</td>\n",
       "      <td>0.970511</td>\n",
       "      <td>-0.334307</td>\n",
       "      <td>-0.633618</td>\n",
       "      <td>0.947049</td>\n",
       "      <td>-0.501684</td>\n",
       "      <td>0.861559</td>\n",
       "      <td>0.052578</td>\n",
       "      <td>-0.321685</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.274229</td>\n",
       "      <td>-0.674483</td>\n",
       "      <td>0.109402</td>\n",
       "      <td>0.601827</td>\n",
       "      <td>-0.480361</td>\n",
       "      <td>0.092184</td>\n",
       "      <td>0.221266</td>\n",
       "      <td>0.075206</td>\n",
       "      <td>4.56</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9457</td>\n",
       "      <td>13993.0</td>\n",
       "      <td>1.085876</td>\n",
       "      <td>-0.048652</td>\n",
       "      <td>1.275049</td>\n",
       "      <td>1.307756</td>\n",
       "      <td>-0.519074</td>\n",
       "      <td>0.827831</td>\n",
       "      <td>-0.846353</td>\n",
       "      <td>0.325183</td>\n",
       "      <td>1.958550</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.154765</td>\n",
       "      <td>-0.007733</td>\n",
       "      <td>-0.017300</td>\n",
       "      <td>-0.362792</td>\n",
       "      <td>0.326210</td>\n",
       "      <td>-0.415107</td>\n",
       "      <td>0.056283</td>\n",
       "      <td>0.015306</td>\n",
       "      <td>12.99</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>92672</td>\n",
       "      <td>64040.0</td>\n",
       "      <td>-0.894390</td>\n",
       "      <td>0.520338</td>\n",
       "      <td>2.032163</td>\n",
       "      <td>-0.034288</td>\n",
       "      <td>0.746082</td>\n",
       "      <td>-0.447069</td>\n",
       "      <td>0.526433</td>\n",
       "      <td>0.015126</td>\n",
       "      <td>-0.735332</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.165533</td>\n",
       "      <td>-0.523333</td>\n",
       "      <td>-0.090716</td>\n",
       "      <td>0.164864</td>\n",
       "      <td>-0.035810</td>\n",
       "      <td>0.100132</td>\n",
       "      <td>-0.152945</td>\n",
       "      <td>-0.148412</td>\n",
       "      <td>2.68</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>240674</td>\n",
       "      <td>150703.0</td>\n",
       "      <td>-0.101805</td>\n",
       "      <td>1.058591</td>\n",
       "      <td>0.174630</td>\n",
       "      <td>0.967532</td>\n",
       "      <td>1.069017</td>\n",
       "      <td>-0.869342</td>\n",
       "      <td>1.546421</td>\n",
       "      <td>-0.471847</td>\n",
       "      <td>-0.419273</td>\n",
       "      <td>...</td>\n",
       "      <td>0.066509</td>\n",
       "      <td>0.638260</td>\n",
       "      <td>-0.293054</td>\n",
       "      <td>-0.012446</td>\n",
       "      <td>-0.014162</td>\n",
       "      <td>-0.401708</td>\n",
       "      <td>0.267594</td>\n",
       "      <td>0.021102</td>\n",
       "      <td>14.80</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>213467</td>\n",
       "      <td>139269.0</td>\n",
       "      <td>2.120229</td>\n",
       "      <td>-0.188559</td>\n",
       "      <td>-1.859070</td>\n",
       "      <td>-0.203124</td>\n",
       "      <td>0.495329</td>\n",
       "      <td>-0.506593</td>\n",
       "      <td>0.175342</td>\n",
       "      <td>-0.143356</td>\n",
       "      <td>0.654438</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.176695</td>\n",
       "      <td>-0.404786</td>\n",
       "      <td>0.059672</td>\n",
       "      <td>-1.116987</td>\n",
       "      <td>0.107433</td>\n",
       "      <td>-0.026732</td>\n",
       "      <td>-0.062180</td>\n",
       "      <td>-0.086943</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Time        V1        V2        V3        V4        V5        V6  \\\n",
       "186521  127145.0  0.077780  0.970511 -0.334307 -0.633618  0.947049 -0.501684   \n",
       "9457     13993.0  1.085876 -0.048652  1.275049  1.307756 -0.519074  0.827831   \n",
       "92672    64040.0 -0.894390  0.520338  2.032163 -0.034288  0.746082 -0.447069   \n",
       "240674  150703.0 -0.101805  1.058591  0.174630  0.967532  1.069017 -0.869342   \n",
       "213467  139269.0  2.120229 -0.188559 -1.859070 -0.203124  0.495329 -0.506593   \n",
       "\n",
       "              V7        V8        V9  ...       V21       V22       V23  \\\n",
       "186521  0.861559  0.052578 -0.321685  ... -0.274229 -0.674483  0.109402   \n",
       "9457   -0.846353  0.325183  1.958550  ... -0.154765 -0.007733 -0.017300   \n",
       "92672   0.526433  0.015126 -0.735332  ... -0.165533 -0.523333 -0.090716   \n",
       "240674  1.546421 -0.471847 -0.419273  ...  0.066509  0.638260 -0.293054   \n",
       "213467  0.175342 -0.143356  0.654438  ... -0.176695 -0.404786  0.059672   \n",
       "\n",
       "             V24       V25       V26       V27       V28  Amount  Class  \n",
       "186521  0.601827 -0.480361  0.092184  0.221266  0.075206    4.56      0  \n",
       "9457   -0.362792  0.326210 -0.415107  0.056283  0.015306   12.99      0  \n",
       "92672   0.164864 -0.035810  0.100132 -0.152945 -0.148412    2.68      0  \n",
       "240674 -0.012446 -0.014162 -0.401708  0.267594  0.021102   14.80      0  \n",
       "213467 -1.116987  0.107433 -0.026732 -0.062180 -0.086943    0.00      0  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# reading in CSV\n",
    "df = pd.read_csv('creditcard.csv')\n",
    "df.sample(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Time', 'V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9',\n",
       "       'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18',\n",
       "       'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27',\n",
       "       'V28', 'Amount', 'Class'], dtype=object)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount', 'Class']\n",
      "SMOTE Oversampling selected..\n",
      "Number of Xs and Ys for: SMOTE\n",
      "[(0, 227451), (1, 227451)]\n",
      "Oversampling is complete!\n",
      "i is 0\n",
      "j is 0\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "()\n",
      "False\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.575227\n",
      "         Iterations 7\n",
      "First model trained using feature number 0 with p value of 0\n",
      "Full X:\n",
      "(454902, 1)\n",
      "i is 1\n",
      "j is 1\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 1)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.512057\n",
      "         Iterations 7\n",
      "Features all have p value of 0, using feature number 1\n",
      "i is 2\n",
      "j is 2\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 2)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.449886\n",
      "         Iterations 8\n",
      "Features all have p value of 0, using feature number 2\n",
      "i is 3\n",
      "j is 3\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 3)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.421939\n",
      "         Iterations 8\n",
      "Features all have p value of 0, using feature number 3\n",
      "i is 4\n",
      "j is 4\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 4)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420662\n",
      "         Iterations 9\n",
      "i is 4\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 4)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.420621\n",
      "         Iterations 8\n",
      "i is 4\n",
      "j is 6\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 4)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.415746\n",
      "         Iterations 9\n",
      "Features all have p value of 0, using feature number 6\n",
      "i is 5\n",
      "j is 4\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.414780\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.414587\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 7\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.415711\n",
      "         Iterations 9\n",
      "i is 5\n",
      "j is 8\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 5)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412989\n",
      "         Iterations 9\n",
      "Features all have p value of 0, using feature number 8\n",
      "i is 6\n",
      "j is 4\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412132\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412040\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 7\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412858\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.401061\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 10\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.405644\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 11\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.399294\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412863\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 13\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.389119\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412988\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 15\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409330\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 16\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409572\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412614\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412946\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 19\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412929\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412366\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412977\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 22\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412909\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412850\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412775\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412903\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 26\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412758\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.412833\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 28\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.377994\n",
      "         Iterations 9\n",
      "i is 6\n",
      "j is 29\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 6)\n",
      "True\n",
      "Warning: Maximum number of iterations has been exceeded.\n",
      "         Current function value: 0.346499\n",
      "         Iterations: 35\n",
      "Getting new model with min p-values with 6 variables.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py:512: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  \"Check mle_retvals\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "New model trained using feature number 15 with lowest p values of 8.21550214335162e-255\n",
      "i is 7\n",
      "j is 4\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409110\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.408413\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 7\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409141\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.399576\n",
      "         Iterations 10\n",
      "i is 7\n",
      "j is 10\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.403039\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 11\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.397918\n",
      "         Iterations 10\n",
      "i is 7\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409174\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 13\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.387728\n",
      "         Iterations 10\n",
      "i is 7\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409319\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 16\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.408123\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409321\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.408999\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 19\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409313\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.408840\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409326\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 22\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409221\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409152\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409041\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409229\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 26\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409122\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.409191\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 28\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.368584\n",
      "         Iterations 9\n",
      "i is 7\n",
      "j is 29\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 7)\n",
      "True\n",
      "Warning: Maximum number of iterations has been exceeded.\n",
      "         Current function value: 0.346465\n",
      "         Iterations: 35\n",
      "Getting new model with min p-values with 7 variables.\n",
      "New model trained using feature number 28 with lowest p values of 9.022727667572669e-241\n",
      "i is 8\n",
      "j is 4\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 8)\n",
      "True"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py:512: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  \"Check mle_retvals\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.346687\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 4\n",
      "i is 9\n",
      "j is 5\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 9)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.344718\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 5\n",
      "i is 10\n",
      "j is 7\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 10)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.338794\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 7\n",
      "i is 11\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 11)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.291590\n",
      "         Iterations 10\n",
      "i is 11\n",
      "j is 10\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 11)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.323632\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 10\n",
      "i is 12\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 12)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.286201\n",
      "         Iterations 10\n",
      "i is 12\n",
      "j is 11\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 12)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.305176\n",
      "         Iterations 10\n",
      "i is 12\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 12)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.323631\n",
      "         Iterations 10\n",
      "i is 12\n",
      "j is 13\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 12)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.292196\n",
      "         Iterations 10\n",
      "Features all have p value of 0, using feature number 13\n",
      "i is 13\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 13)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.273086\n",
      "         Iterations 10\n",
      "i is 13\n",
      "j is 11\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 13)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.276701\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 11\n",
      "i is 14\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.255552\n",
      "         Iterations 11\n",
      "i is 14\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.276572\n",
      "         Iterations 11\n",
      "i is 14\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.276573\n",
      "         Iterations 11\n",
      "i is 14\n",
      "j is 16\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.271407\n",
      "         Iterations 12\n",
      "i is 14\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.276501\n",
      "         Iterations 11\n",
      "i is 14\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.276051\n",
      "         Iterations 11\n",
      "i is 14\n",
      "j is 19\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 14)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.262932\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 19\n",
      "i is 15\n",
      "j is 9\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 15)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.238806\n",
      "         Iterations 11\n",
      "Features all have p value of 0, using feature number 9\n",
      "i is 16\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 16)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.238798\n",
      "         Iterations 11\n",
      "i is 16\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 16)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.238767\n",
      "         Iterations 11\n",
      "i is 16\n",
      "j is 16\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 16)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.224329\n",
      "         Iterations 13\n",
      "Features all have p value of 0, using feature number 16\n",
      "i is 17\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.224312\n",
      "         Iterations 13\n",
      "i is 17\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.224213\n",
      "         Iterations 13\n",
      "i is 17\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.220808\n",
      "         Iterations 14\n",
      "i is 17\n",
      "j is 18\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 17)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.222467\n",
      "         Iterations 13\n",
      "Features all have p value of 0, using feature number 18\n",
      "i is 18\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.222463\n",
      "         Iterations 13\n",
      "i is 18\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.222428\n",
      "         Iterations 13\n",
      "i is 18\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.220081\n",
      "         Iterations 14\n",
      "i is 18\n",
      "j is 20\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 18)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210426\n",
      "         Iterations 13\n",
      "Features all have p value of 0, using feature number 20\n",
      "i is 19\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210413\n",
      "         Iterations 13\n",
      "i is 19\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.210360\n",
      "         Iterations 13\n",
      "i is 19\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.205460\n",
      "         Iterations 15\n",
      "i is 19\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.208885\n",
      "         Iterations 13\n",
      "i is 19\n",
      "j is 22\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 19)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.190005\n",
      "         Iterations 14\n",
      "Features all have p value of 0, using feature number 22\n",
      "i is 20\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.189952\n",
      "         Iterations 14\n",
      "i is 20\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.189995\n",
      "         Iterations 14\n",
      "i is 20\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.188009\n",
      "         Iterations 16\n",
      "i is 20\n",
      "j is 21\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 20)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.186116\n",
      "         Iterations 14\n",
      "Features all have p value of 0, using feature number 21\n",
      "i is 21\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.186088\n",
      "         Iterations 14\n",
      "i is 21\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.186101\n",
      "         Iterations 14\n",
      "i is 21\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.184392\n",
      "         Iterations 15\n",
      "i is 21\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.186112\n",
      "         Iterations 14\n",
      "i is 21\n",
      "j is 24\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 21)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.180247\n",
      "         Iterations 14\n",
      "Features all have p value of 0, using feature number 24\n",
      "i is 22\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.180236\n",
      "         Iterations 14\n",
      "i is 22\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.180219\n",
      "         Iterations 14\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "i is 22\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.179266\n",
      "         Iterations 16\n",
      "i is 22\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.180244\n",
      "         Iterations 14\n",
      "i is 22\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.179832\n",
      "         Iterations 14\n",
      "i is 22\n",
      "j is 26\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 22)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.177032\n",
      "         Iterations 14\n",
      "Features all have p value of 0, using feature number 26\n",
      "i is 23\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.177017\n",
      "         Iterations 14\n",
      "i is 23\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.176982\n",
      "         Iterations 14\n",
      "i is 23\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.176130\n",
      "         Iterations 16\n",
      "i is 23\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.177032\n",
      "         Iterations 14\n",
      "i is 23\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.176796\n",
      "         Iterations 14\n",
      "i is 23\n",
      "j is 27\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 23)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174490\n",
      "         Iterations 15\n",
      "Features all have p value of 0, using feature number 27\n",
      "i is 24\n",
      "j is 12\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174483\n",
      "         Iterations 15\n",
      "i is 24\n",
      "j is 14\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174456\n",
      "         Iterations 15\n",
      "i is 24\n",
      "j is 17\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.173430\n",
      "         Iterations 16\n",
      "i is 24\n",
      "j is 23\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174489\n",
      "         Iterations 15\n",
      "i is 24\n",
      "j is 25\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 24)\n",
      "True\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.174223\n",
      "         Iterations 15\n",
      "i is 24\n",
      "j is 29\n",
      "Jth_x:\n",
      "(454902, 1)\n",
      "Full_x:\n",
      "(454902, 24)\n",
      "True\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\discrete\\discrete_model.py:1789: RuntimeWarning: divide by zero encountered in log\n",
      "  return np.sum(np.log(self.cdf(q*np.dot(X,params))))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Warning: Maximum number of iterations has been exceeded.\n",
      "         Current function value: inf\n",
      "         Iterations: 35\n"
     ]
    },
    {
     "ename": "LinAlgError",
     "evalue": "Singular matrix",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mLinAlgError\u001b[0m                               Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-25-a58734844a14>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mfinalcolstouse\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mforwardselection\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdf\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;36m0.2\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'Class'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'Time'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'smote'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32m<ipython-input-22-1d76b6a05ec9>\u001b[0m in \u001b[0;36mforwardselection\u001b[1;34m(df, testratio, response, removelist, sampling)\u001b[0m\n\u001b[0;32m     85\u001b[0m                 \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     86\u001b[0m                     \u001b[0miter_x\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mjth_x\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 87\u001b[1;33m                 \u001b[0mregressor_OLS\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0msm\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mLogit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0my_train\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0miter_x\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\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     88\u001b[0m                 \u001b[1;31m# print(\"p value for column number: \" + str(i))\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     89\u001b[0m                 \u001b[1;31m# print(regressor_OLS.pvalues[0])\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\statsmodels\\discrete\\discrete_model.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, start_params, method, maxiter, full_output, disp, callback, **kwargs)\u001b[0m\n\u001b[0;32m   1900\u001b[0m         bnryfit = super(Logit, self).fit(start_params=start_params,\n\u001b[0;32m   1901\u001b[0m                 \u001b[0mmethod\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mmethod\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmaxiter\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mmaxiter\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfull_output\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mfull_output\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1902\u001b[1;33m                 disp=disp, callback=callback, **kwargs)\n\u001b[0m\u001b[0;32m   1903\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1904\u001b[0m         \u001b[0mdiscretefit\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mLogitResults\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mbnryfit\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\statsmodels\\discrete\\discrete_model.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, start_params, method, maxiter, full_output, disp, callback, **kwargs)\u001b[0m\n\u001b[0;32m    215\u001b[0m         mlefit = super(DiscreteModel, self).fit(start_params=start_params,\n\u001b[0;32m    216\u001b[0m                 \u001b[0mmethod\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mmethod\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmaxiter\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mmaxiter\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfull_output\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mfull_output\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 217\u001b[1;33m                 disp=disp, callback=callback, **kwargs)\n\u001b[0m\u001b[0;32m    218\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    219\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mmlefit\u001b[0m \u001b[1;31m# up to subclasses to wrap results\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, start_params, method, maxiter, full_output, disp, fargs, callback, retall, skip_hessian, **kwargs)\u001b[0m\n\u001b[0;32m    475\u001b[0m             \u001b[0mHinv\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mcov_params_func\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mxopt\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mretvals\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    476\u001b[0m         \u001b[1;32melif\u001b[0m \u001b[0mmethod\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;34m'newton'\u001b[0m \u001b[1;32mand\u001b[0m \u001b[0mfull_output\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 477\u001b[1;33m             \u001b[0mHinv\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mlinalg\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minv\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m-\u001b[0m\u001b[0mretvals\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'Hessian'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m/\u001b[0m \u001b[0mnobs\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    478\u001b[0m         \u001b[1;32melif\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[0mskip_hessian\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    479\u001b[0m             \u001b[0mH\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m-\u001b[0m\u001b[1;36m1\u001b[0m \u001b[1;33m*\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mhessian\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mxopt\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\numpy\\linalg\\linalg.py\u001b[0m in \u001b[0;36minv\u001b[1;34m(a)\u001b[0m\n\u001b[0;32m    549\u001b[0m     \u001b[0msignature\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;34m'D->D'\u001b[0m \u001b[1;32mif\u001b[0m \u001b[0misComplexType\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mt\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32melse\u001b[0m \u001b[1;34m'd->d'\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    550\u001b[0m     \u001b[0mextobj\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mget_linalg_error_extobj\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0m_raise_linalgerror_singular\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 551\u001b[1;33m     \u001b[0mainv\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0m_umath_linalg\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minv\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0ma\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0msignature\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0msignature\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mextobj\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mextobj\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    552\u001b[0m     \u001b[1;32mreturn\u001b[0m \u001b[0mwrap\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mainv\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mastype\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mresult_t\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcopy\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mFalse\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    553\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\numpy\\linalg\\linalg.py\u001b[0m in \u001b[0;36m_raise_linalgerror_singular\u001b[1;34m(err, flag)\u001b[0m\n\u001b[0;32m     95\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     96\u001b[0m \u001b[1;32mdef\u001b[0m \u001b[0m_raise_linalgerror_singular\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0merr\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mflag\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 97\u001b[1;33m     \u001b[1;32mraise\u001b[0m \u001b[0mLinAlgError\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"Singular matrix\"\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     98\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     99\u001b[0m \u001b[1;32mdef\u001b[0m \u001b[0m_raise_linalgerror_nonposdef\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0merr\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mflag\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mLinAlgError\u001b[0m: Singular matrix"
     ]
    }
   ],
   "source": [
    "finalcolstouse = forwardselection(df,0.2,'Class','Time','smote')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Forward Selection Failed\n",
    "## Could be because after oversampling, there is high collinearity and relation between features, leading to singular matrix calc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Backward Elimination Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "import statsmodels.discrete.discrete_model as sm\n",
    "def backwardElimination(df , sl, columns, response, sampling):\n",
    "    df = pd.concat([df[columns],df[response]], axis=1)\n",
    "    x = df[columns].values\n",
    "    Y = df[response].values\n",
    "    \n",
    "    # Sampling if required\n",
    "    if sampling.lower() == 'smote':\n",
    "        print(\"SMOTE Oversampling selected..\")\n",
    "        x, Y = SMOTE().fit_resample(x, Y)\n",
    "        # train test split keeps X_test and y_test as pd series, oversampler converts X_train, y_train to numpy\n",
    "        # Convert all to numpy array for XGBoost to not have bugs\n",
    "        print(\"Number of Xs and Ys for: \" + str(sampling.upper()))\n",
    "        print(sorted(Counter(Y).items()))\n",
    "        print(\"Oversampling is complete!\")\n",
    "    elif sampling.lower() == 'naive':\n",
    "        print(\"Naive Oversampling selected..\")\n",
    "        ros = RandomOverSampler(random_state=42)\n",
    "        x, Y = ros.fit_resample(x, Y)\n",
    "        print(\"Number of Xs and Ys for: \" + str(sampling.upper()))\n",
    "        print(sorted(Counter(Y).items()))\n",
    "        print(\"Oversampling is complete!\")\n",
    "    \n",
    "    numVars = len(x[0]) # Get length of a row for num of vars\n",
    "    for i in range(0, numVars):\n",
    "        sm_Logit = sm.Logit(Y, x).fit() \n",
    "        # for loop to fit current set of all vars in x and response Y\n",
    "        # As loop goes on, x gets smaller as columns are deleted, and columns keeping track of col names also edited\n",
    "        maxVar = max(sm_Logit.pvalues).astype(float)\n",
    "        print('Regression model retrained with ' + str(numVars) + ' variables.')\n",
    "        print('Max p value for a feature is: ' + str(maxVar))\n",
    "        # get max p value and if its more than sig level, start deleting jth column\n",
    "        # Since columns are getting deleted and x gets smaller, we need to update columns keeping track of column names\n",
    "        # Hence the only way to ensure the deletion of right column\n",
    "        # is to check the max p value with the current pvalues[j] in regression model\n",
    "        # if they are same, then jth column safe to delete\n",
    "        if maxVar > sl:\n",
    "            print('Max p value > ' + str(sl) + ', feature will be removed.')\n",
    "            for j in range(0, numVars - i):\n",
    "                if (sm_Logit.pvalues[j].astype(float) == maxVar):\n",
    "                    print(str(j) + 'th column deleted: ' + str(columns[j]))\n",
    "                    x = np.delete(x, j, 1)\n",
    "                    columns = np.delete(columns, j)\n",
    "                    numVars -= 1\n",
    "        else:\n",
    "            print('All p values are above ' + str(sl) + '. Terminating model training')\n",
    "            print('p values list: ' + str(sm_Logit.pvalues))\n",
    "            break\n",
    "    \n",
    "    # Only in case df needs to be output\n",
    "    # df = pd.concat([pd.DataFrame(x, columns=columns), df[response]], axis=1)\n",
    "    print(sm_Logit.summary())\n",
    "    return columns # Return x data and list of columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount']\n"
     ]
    }
   ],
   "source": [
    "allfeatures = list(df.columns.values)\n",
    "allfeatures.remove('Class')\n",
    "allfeatures.remove('Time')\n",
    "print(allfeatures)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "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>Time</th>\n",
       "      <th>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>V9</th>\n",
       "      <th>...</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.359807</td>\n",
       "      <td>-0.072781</td>\n",
       "      <td>2.536347</td>\n",
       "      <td>1.378155</td>\n",
       "      <td>-0.338321</td>\n",
       "      <td>0.462388</td>\n",
       "      <td>0.239599</td>\n",
       "      <td>0.098698</td>\n",
       "      <td>0.363787</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.018307</td>\n",
       "      <td>0.277838</td>\n",
       "      <td>-0.110474</td>\n",
       "      <td>0.066928</td>\n",
       "      <td>0.128539</td>\n",
       "      <td>-0.189115</td>\n",
       "      <td>0.133558</td>\n",
       "      <td>-0.021053</td>\n",
       "      <td>149.62</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.191857</td>\n",
       "      <td>0.266151</td>\n",
       "      <td>0.166480</td>\n",
       "      <td>0.448154</td>\n",
       "      <td>0.060018</td>\n",
       "      <td>-0.082361</td>\n",
       "      <td>-0.078803</td>\n",
       "      <td>0.085102</td>\n",
       "      <td>-0.255425</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.225775</td>\n",
       "      <td>-0.638672</td>\n",
       "      <td>0.101288</td>\n",
       "      <td>-0.339846</td>\n",
       "      <td>0.167170</td>\n",
       "      <td>0.125895</td>\n",
       "      <td>-0.008983</td>\n",
       "      <td>0.014724</td>\n",
       "      <td>2.69</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.358354</td>\n",
       "      <td>-1.340163</td>\n",
       "      <td>1.773209</td>\n",
       "      <td>0.379780</td>\n",
       "      <td>-0.503198</td>\n",
       "      <td>1.800499</td>\n",
       "      <td>0.791461</td>\n",
       "      <td>0.247676</td>\n",
       "      <td>-1.514654</td>\n",
       "      <td>...</td>\n",
       "      <td>0.247998</td>\n",
       "      <td>0.771679</td>\n",
       "      <td>0.909412</td>\n",
       "      <td>-0.689281</td>\n",
       "      <td>-0.327642</td>\n",
       "      <td>-0.139097</td>\n",
       "      <td>-0.055353</td>\n",
       "      <td>-0.059752</td>\n",
       "      <td>378.66</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.966272</td>\n",
       "      <td>-0.185226</td>\n",
       "      <td>1.792993</td>\n",
       "      <td>-0.863291</td>\n",
       "      <td>-0.010309</td>\n",
       "      <td>1.247203</td>\n",
       "      <td>0.237609</td>\n",
       "      <td>0.377436</td>\n",
       "      <td>-1.387024</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.108300</td>\n",
       "      <td>0.005274</td>\n",
       "      <td>-0.190321</td>\n",
       "      <td>-1.175575</td>\n",
       "      <td>0.647376</td>\n",
       "      <td>-0.221929</td>\n",
       "      <td>0.062723</td>\n",
       "      <td>0.061458</td>\n",
       "      <td>123.50</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.158233</td>\n",
       "      <td>0.877737</td>\n",
       "      <td>1.548718</td>\n",
       "      <td>0.403034</td>\n",
       "      <td>-0.407193</td>\n",
       "      <td>0.095921</td>\n",
       "      <td>0.592941</td>\n",
       "      <td>-0.270533</td>\n",
       "      <td>0.817739</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.009431</td>\n",
       "      <td>0.798278</td>\n",
       "      <td>-0.137458</td>\n",
       "      <td>0.141267</td>\n",
       "      <td>-0.206010</td>\n",
       "      <td>0.502292</td>\n",
       "      <td>0.219422</td>\n",
       "      <td>0.215153</td>\n",
       "      <td>69.99</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Time        V1        V2        V3        V4        V5        V6        V7  \\\n",
       "0   0.0 -1.359807 -0.072781  2.536347  1.378155 -0.338321  0.462388  0.239599   \n",
       "1   0.0  1.191857  0.266151  0.166480  0.448154  0.060018 -0.082361 -0.078803   \n",
       "2   1.0 -1.358354 -1.340163  1.773209  0.379780 -0.503198  1.800499  0.791461   \n",
       "3   1.0 -0.966272 -0.185226  1.792993 -0.863291 -0.010309  1.247203  0.237609   \n",
       "4   2.0 -1.158233  0.877737  1.548718  0.403034 -0.407193  0.095921  0.592941   \n",
       "\n",
       "         V8        V9  ...       V21       V22       V23       V24       V25  \\\n",
       "0  0.098698  0.363787  ... -0.018307  0.277838 -0.110474  0.066928  0.128539   \n",
       "1  0.085102 -0.255425  ... -0.225775 -0.638672  0.101288 -0.339846  0.167170   \n",
       "2  0.247676 -1.514654  ...  0.247998  0.771679  0.909412 -0.689281 -0.327642   \n",
       "3  0.377436 -1.387024  ... -0.108300  0.005274 -0.190321 -1.175575  0.647376   \n",
       "4 -0.270533  0.817739  ... -0.009431  0.798278 -0.137458  0.141267 -0.206010   \n",
       "\n",
       "        V26       V27       V28  Amount  Class  \n",
       "0 -0.189115  0.133558 -0.021053  149.62      0  \n",
       "1  0.125895 -0.008983  0.014724    2.69      0  \n",
       "2 -0.139097 -0.055353 -0.059752  378.66      0  \n",
       "3 -0.221929  0.062723  0.061458  123.50      0  \n",
       "4  0.502292  0.219422  0.215153   69.99      0  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Naive Oversampling selected..\n",
      "Number of Xs and Ys for: NAIVE\n",
      "[(0, 284315), (1, 284315)]\n",
      "Oversampling is complete!\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212561\n",
      "         Iterations 14\n",
      "Regression model retrained with 29 variables.\n",
      "Max p value for a feature is: 0.20333745292347039\n",
      "Max p value > 0.05, feature will be removed.\n",
      "17th column deleted: V18\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.212563\n",
      "         Iterations 13\n",
      "Regression model retrained with 28 variables.\n",
      "Max p value for a feature is: 7.846441999162881e-52\n",
      "All p values are above 0.05. Terminating model training\n",
      "p values list: [0.00000000e+000 0.00000000e+000 0.00000000e+000 0.00000000e+000\n",
      " 0.00000000e+000 0.00000000e+000 0.00000000e+000 0.00000000e+000\n",
      " 0.00000000e+000 0.00000000e+000 0.00000000e+000 0.00000000e+000\n",
      " 1.29758484e-124 0.00000000e+000 5.32212829e-070 0.00000000e+000\n",
      " 0.00000000e+000 0.00000000e+000 0.00000000e+000 0.00000000e+000\n",
      " 0.00000000e+000 0.00000000e+000 7.84644200e-052 0.00000000e+000\n",
      " 1.16923494e-186 0.00000000e+000 0.00000000e+000 0.00000000e+000]\n",
      "                           Logit Regression Results                           \n",
      "==============================================================================\n",
      "Dep. Variable:                      y   No. Observations:               568630\n",
      "Model:                          Logit   Df Residuals:                   568602\n",
      "Method:                           MLE   Df Model:                           27\n",
      "Date:                Thu, 06 Feb 2020   Pseudo R-squ.:                  0.6933\n",
      "Time:                        16:56:28   Log-Likelihood:            -1.2087e+05\n",
      "converged:                       True   LL-Null:                   -3.9414e+05\n",
      "Covariance Type:            nonrobust   LLR p-value:                     0.000\n",
      "==============================================================================\n",
      "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "x1            -0.7759      0.005   -159.769      0.000      -0.785      -0.766\n",
      "x2            -2.8173      0.013   -217.097      0.000      -2.843      -2.792\n",
      "x3            -1.0330      0.005   -201.280      0.000      -1.043      -1.023\n",
      "x4             1.2583      0.006    226.254      0.000       1.247       1.269\n",
      "x5            -1.6347      0.007   -241.303      0.000      -1.648      -1.621\n",
      "x6             0.9321      0.006    155.761      0.000       0.920       0.944\n",
      "x7             2.2958      0.011    216.553      0.000       2.275       2.317\n",
      "x8            -1.1970      0.008   -156.793      0.000      -1.212      -1.182\n",
      "x9            -0.4450      0.007    -67.589      0.000      -0.458      -0.432\n",
      "x10           -1.1569      0.009   -123.110      0.000      -1.175      -1.138\n",
      "x11            0.5439      0.006     87.258      0.000       0.532       0.556\n",
      "x12           -0.9669      0.008   -120.576      0.000      -0.983      -0.951\n",
      "x13           -0.1208      0.005    -23.743      0.000      -0.131      -0.111\n",
      "x14           -0.8985      0.008   -111.936      0.000      -0.914      -0.883\n",
      "x15            0.0986      0.006     17.687      0.000       0.088       0.110\n",
      "x16           -0.4278      0.008    -51.789      0.000      -0.444      -0.412\n",
      "x17           -0.6861      0.010    -65.893      0.000      -0.706      -0.666\n",
      "x18           -0.3332      0.006    -53.095      0.000      -0.345      -0.321\n",
      "x19            2.1179      0.011    188.397      0.000       2.096       2.140\n",
      "x20            0.8077      0.007    110.848      0.000       0.793       0.822\n",
      "x21           -0.4993      0.008    -65.529      0.000      -0.514      -0.484\n",
      "x22           -2.4334      0.010   -232.778      0.000      -2.454      -2.413\n",
      "x23            0.1435      0.009     15.148      0.000       0.125       0.162\n",
      "x24           -0.9113      0.011    -83.321      0.000      -0.933      -0.890\n",
      "x25           -0.3372      0.012    -29.138      0.000      -0.360      -0.315\n",
      "x26            1.2487      0.019     66.144      0.000       1.212       1.286\n",
      "x27            1.0868      0.027     39.567      0.000       1.033       1.141\n",
      "x28           -0.0256   9.15e-05   -279.287      0.000      -0.026      -0.025\n",
      "==============================================================================\n",
      "\n",
      "Possibly complete quasi-separation: A fraction 0.40 of observations can be\n",
      "perfectly predicted. This might indicate that there is complete\n",
      "quasi-separation. In this case some parameters will not be identified.\n"
     ]
    }
   ],
   "source": [
    "selectedcols = backwardElimination(df , 0.05, allfeatures, 'Class', 'naive')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1' 'V2' 'V3' 'V4' 'V5' 'V6' 'V7' 'V8' 'V9' 'V10' 'V11' 'V12' 'V13'\n",
      " 'V14' 'V15' 'V16' 'V17' 'V19' 'V20' 'V21' 'V22' 'V23' 'V24' 'V25' 'V26'\n",
      " 'V27' 'V28' 'Amount']\n"
     ]
    }
   ],
   "source": [
    "print(selectedcols)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SMOTE Oversampling selected..\n",
      "Number of Xs and Ys for: SMOTE\n",
      "[(0, 284315), (1, 284315)]\n",
      "Oversampling is complete!\n",
      "Optimization terminated successfully.\n",
      "         Current function value: 0.176939\n",
      "         Iterations 15\n",
      "Regression model retrained with 29 variables.\n",
      "Max p value for a feature is: 0.005680664995002565\n",
      "All p values are above 0.05. Terminating model training\n",
      "p values list: [0.00000000e+000 0.00000000e+000 0.00000000e+000 0.00000000e+000\n",
      " 0.00000000e+000 0.00000000e+000 0.00000000e+000 0.00000000e+000\n",
      " 0.00000000e+000 0.00000000e+000 0.00000000e+000 0.00000000e+000\n",
      " 3.72529026e-044 0.00000000e+000 1.98842391e-030 0.00000000e+000\n",
      " 0.00000000e+000 5.72602538e-183 0.00000000e+000 0.00000000e+000\n",
      " 0.00000000e+000 0.00000000e+000 0.00000000e+000 5.68066500e-003\n",
      " 0.00000000e+000 6.46118928e-039 0.00000000e+000 0.00000000e+000\n",
      " 0.00000000e+000]\n",
      "                           Logit Regression Results                           \n",
      "==============================================================================\n",
      "Dep. Variable:                      y   No. Observations:               568630\n",
      "Model:                          Logit   Df Residuals:                   568601\n",
      "Method:                           MLE   Df Model:                           28\n",
      "Date:                Thu, 06 Feb 2020   Pseudo R-squ.:                  0.7447\n",
      "Time:                        16:56:51   Log-Likelihood:            -1.0061e+05\n",
      "converged:                       True   LL-Null:                   -3.9414e+05\n",
      "Covariance Type:            nonrobust   LLR p-value:                     0.000\n",
      "==============================================================================\n",
      "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "x1            -0.8482      0.006   -135.882      0.000      -0.860      -0.836\n",
      "x2            -2.6489      0.016   -162.276      0.000      -2.681      -2.617\n",
      "x3            -1.3601      0.009   -143.479      0.000      -1.379      -1.342\n",
      "x4             1.4251      0.007    200.581      0.000       1.411       1.439\n",
      "x5            -1.9849      0.009   -221.358      0.000      -2.002      -1.967\n",
      "x6             0.9407      0.008    120.927      0.000       0.925       0.956\n",
      "x7             1.9022      0.019     98.950      0.000       1.865       1.940\n",
      "x8            -1.0646      0.010   -107.265      0.000      -1.084      -1.045\n",
      "x9            -0.7501      0.009    -85.639      0.000      -0.767      -0.733\n",
      "x10           -1.6317      0.015   -110.970      0.000      -1.660      -1.603\n",
      "x11            0.9163      0.010     87.397      0.000       0.896       0.937\n",
      "x12           -1.5723      0.017    -93.187      0.000      -1.605      -1.539\n",
      "x13           -0.0814      0.006    -13.938      0.000      -0.093      -0.070\n",
      "x14           -1.5090      0.017    -86.700      0.000      -1.543      -1.475\n",
      "x15            0.0736      0.006     11.465      0.000       0.061       0.086\n",
      "x16           -1.0169      0.017    -60.447      0.000      -1.050      -0.984\n",
      "x17           -1.6714      0.027    -60.939      0.000      -1.725      -1.618\n",
      "x18           -0.3619      0.013    -28.846      0.000      -0.386      -0.337\n",
      "x19           -0.3548      0.008    -45.114      0.000      -0.370      -0.339\n",
      "x20            2.8668      0.017    172.283      0.000       2.834       2.899\n",
      "x21            1.0805      0.009    119.573      0.000       1.063       1.098\n",
      "x22           -0.5796      0.009    -64.053      0.000      -0.597      -0.562\n",
      "x23           -2.2801      0.017   -133.705      0.000      -2.314      -2.247\n",
      "x24            0.0296      0.011      2.766      0.006       0.009       0.051\n",
      "x25           -0.7953      0.013    -61.521      0.000      -0.821      -0.770\n",
      "x26           -0.1738      0.013    -13.049      0.000      -0.200      -0.148\n",
      "x27            1.0389      0.024     42.517      0.000       0.991       1.087\n",
      "x28            1.7292      0.031     55.319      0.000       1.668       1.790\n",
      "x29           -0.0291      0.000   -269.951      0.000      -0.029      -0.029\n",
      "==============================================================================\n",
      "\n",
      "Possibly complete quasi-separation: A fraction 0.42 of observations can be\n",
      "perfectly predicted. This might indicate that there is complete\n",
      "quasi-separation. In this case some parameters will not be identified.\n"
     ]
    }
   ],
   "source": [
    "selectedcols = backwardElimination(df , 0.05, allfeatures, 'Class', 'smote')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount']\n"
     ]
    }
   ],
   "source": [
    "print(selectedcols)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Forward Selection without comments and with feature name chosen"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Slight improvement to previous code by having more informative print text on what feature name is chosen instead of it's number\n",
    "### Reduction of excessive comments printed out by the function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import operator\n",
    "from collections import Counter\n",
    "from imblearn.over_sampling import RandomOverSampler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from imblearn.over_sampling import ADASYN\n",
    "import statsmodels.discrete.discrete_model as sm\n",
    "\n",
    "def forwardselection(df,testratio,response,removelist,sampling):\n",
    "    if isinstance(removelist, str) == True:\n",
    "        temp_str = removelist\n",
    "        internallist = []\n",
    "        internallist.append(temp_str)\n",
    "    else:\n",
    "        internallist = removelist\n",
    "    X = df.drop(internallist, axis=1)\n",
    "    y = df[response]\n",
    "    # Get list of column names\n",
    "    colnames = list(X.columns.values)\n",
    "    print(colnames)\n",
    "    \n",
    "    # Start of train-test split and oversampling (if relevant)\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=testratio, random_state=42)\n",
    "    if sampling.lower() == 'smote':\n",
    "        print(\"SMOTE Oversampling selected..\")\n",
    "        X_train, y_train = SMOTE().fit_resample(X_train, y_train)\n",
    "        # train test split keeps X_test and y_test as pd series, oversampler converts X_train, y_train to numpy\n",
    "        # Convert all to numpy array for XGBoost to not have bugs\n",
    "        X_test = X_test.values\n",
    "        y_test = y_test.values\n",
    "        print(\"Number of Xs and Ys for: \" + str(sampling.upper()))\n",
    "        print(sorted(Counter(y_train).items()))\n",
    "        print(\"Oversampling is complete!\")\n",
    "    elif sampling.lower() == 'naive':\n",
    "        print(\"Naive Oversampling selected..\")\n",
    "        ros = RandomOverSampler(random_state=42)\n",
    "        X_train, y_train = ros.fit_resample(X_train, y_train)\n",
    "        # train test split keeps X_test and y_test as pd series, oversampler converts X_train, y_train to numpy\n",
    "        # Convert all to numpy array for XGBoost to not have bugs\n",
    "        X_test = X_test.values\n",
    "        y_test = y_test.values\n",
    "        print(\"Number of Xs and Ys for: \" + str(sampling.upper()))\n",
    "        print(sorted(Counter(y_train).items()))\n",
    "        print(\"Oversampling is complete!\")\n",
    "\n",
    "\n",
    "    # Total features to select = k\n",
    "    # In each iteration, the current set of n features is concatenated with a new feature not inside current set\n",
    "    # It is then sent for training with the logistic regression\n",
    "    # The model performance for each feature + current features is evaluated by its highest p value (worst feature)\n",
    "    # All highest p values of all feature addition to n features (k-n iterations) are put into a dictionary\n",
    "    # Next, the lowest p value out of all the iterations (for n features + 1) is chosen for evaluation\n",
    "    # Set significance level, which is compared to the lowest p value of the best model in the current training iteration\n",
    "    # If best model in current training iteration of n vars has any vars with p value > sig level, then the model training stops\n",
    "    # Because all the different models are worse or equally bad as the current best model, we can terminate selection process\n",
    "    # If not, repeat this iteration with now n+1 features and k-n-1 iterations\n",
    "    sig_level = 0.05\n",
    "\n",
    "    maxcolsnum = X_train.shape[1]\n",
    "    full_x = np.array(False)\n",
    "    allowed_nums = {}\n",
    "    for i in range(maxcolsnum):\n",
    "        allowed_nums[i] = True\n",
    "    actual_nums = []\n",
    "    actual_vars = []\n",
    "    terminate_early = False\n",
    "    y = y_train\n",
    "    for i in range(maxcolsnum):\n",
    "        # Reset boolean and pval_list\n",
    "        terminate_early = False\n",
    "        pval_list = {}\n",
    "        for j in range(maxcolsnum):\n",
    "            if allowed_nums[j] == True:\n",
    "                # Need to reshape to single column instead of a long array for concating properly\n",
    "                jth_x = X_train[:,j].reshape(-1,1)\n",
    "                if full_x.any():\n",
    "                    iter_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "                else:\n",
    "                    iter_x = jth_x\n",
    "                regressor_OLS = sm.Logit(y_train, iter_x).fit(disp=0)\n",
    "                pval_list[j] = max(regressor_OLS.pvalues)\n",
    "                # Special condition where all the features have p values of 0, directly use these variables for training\n",
    "                if max(regressor_OLS.pvalues) == 0:\n",
    "                    if full_x.any():\n",
    "                        full_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "                        allowed_nums[j] = False\n",
    "                        actual_nums.append(j)\n",
    "                        print(\"Features all have p value of 0, using feature: [\" + str(colnames[j]) + \"]\")\n",
    "                    else:\n",
    "                        full_x = jth_x\n",
    "                        allowed_nums[j] = False\n",
    "                        actual_nums.append(j)\n",
    "                        print(\"First model trained using feature: [\" + str(colnames[j]) + \"] with p value of 0\")\n",
    "                    terminate_early = True\n",
    "                    break\n",
    "            else:\n",
    "                continue\n",
    "        if i > 0 and terminate_early == False:\n",
    "            print(\"Building new model with lowest p-values with \" + str(len(actual_nums)) + \" variables.\")\n",
    "            max_pval_col = min(pval_list.items(), key=operator.itemgetter(1))[0]\n",
    "            max_pval = pval_list[max_pval_col]\n",
    "            # Need to reshape to single column instead of a long array for concating properly\n",
    "            jth_x = X_train[:,max_pval_col].reshape(-1,1)\n",
    "            if max_pval < sig_level:\n",
    "                if full_x.any():\n",
    "                    full_x = np.concatenate((full_x, jth_x), axis=1)\n",
    "                    allowed_nums[max_pval_col] = False\n",
    "                    actual_nums.append(max_pval_col)\n",
    "                    print(\"New model trained using feature: [\" + str(colnames[max_pval_col]) + \"] with lowest p values of \" + str(max_pval))\n",
    "                else:\n",
    "                    full_x = jth_x\n",
    "                    allowed_nums[max_pval_col] = False\n",
    "                    actual_nums.append(max_pval_col)\n",
    "                    print(\"First model trained using feature: [\" + str(colnames[max_pval_col]) + \"] with lowest p values of \" + str(max_pval))\n",
    "            else:\n",
    "                print(\"TERMINATING AS best model trained using feature: [\" + str(colnames[max_pval_col]) + \"] with high p value of \" + str(max_pval) + \" above significance level: \" + str(sig_level))\n",
    "                break\n",
    "\n",
    "\n",
    "    for k in actual_nums:\n",
    "        actual_vars.append(colnames[k])\n",
    "    print('Final variables selected:')\n",
    "    print(actual_vars)\n",
    "    \n",
    "    return actual_vars"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount', 'Class']\n",
      "Naive Oversampling selected..\n",
      "Number of Xs and Ys for: NAIVE\n",
      "[(0, 255880), (1, 255880)]\n",
      "Oversampling is complete!\n",
      "First model trained using feature: [V1] with p value of 0\n",
      "Features all have p value of 0, using feature: [V2]\n",
      "Features all have p value of 0, using feature: [V3]\n",
      "Features all have p value of 0, using feature: [V4]\n",
      "Features all have p value of 0, using feature: [V6]\n",
      "Features all have p value of 0, using feature: [V7]\n",
      "Features all have p value of 0, using feature: [V17]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py:512: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  \"Check mle_retvals\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Building new model with lowest p-values with 7 variables.\n",
      "New model trained using feature: [V9] with lowest p values of 1.3499052620827247e-291\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py:512: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  \"Check mle_retvals\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Building new model with lowest p-values with 8 variables.\n",
      "New model trained using feature: [V11] with lowest p values of 7.468764810312311e-186\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py:512: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  \"Check mle_retvals\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Building new model with lowest p-values with 9 variables.\n",
      "New model trained using feature: [V8] with lowest p values of 1.7557655661627995e-134\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py:512: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  \"Check mle_retvals\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Building new model with lowest p-values with 10 variables.\n",
      "New model trained using feature: [V16] with lowest p values of 2.025681486168209e-111\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py:512: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  \"Check mle_retvals\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Building new model with lowest p-values with 11 variables.\n",
      "New model trained using feature: [V19] with lowest p values of 5.369924262646178e-88\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py:512: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  \"Check mle_retvals\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Building new model with lowest p-values with 12 variables.\n",
      "New model trained using feature: [V10] with lowest p values of 1.7291020174766805e-66\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py:512: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
      "  \"Check mle_retvals\", ConvergenceWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Building new model with lowest p-values with 13 variables.\n",
      "New model trained using feature: [Amount] with lowest p values of 2.8652322063445345e-133\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\Anaconda3\\lib\\site-packages\\statsmodels\\discrete\\discrete_model.py:1789: RuntimeWarning: divide by zero encountered in log\n",
      "  return np.sum(np.log(self.cdf(q*np.dot(X,params))))\n"
     ]
    },
    {
     "ename": "LinAlgError",
     "evalue": "Singular matrix",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mLinAlgError\u001b[0m                               Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-59-721824ae0937>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mfinalcolstouse\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mforwardselection\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdf\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;36m0.1\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'Class'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'Time'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;34m'naive'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32m<ipython-input-58-4125cdc9c466>\u001b[0m in \u001b[0;36mforwardselection\u001b[1;34m(df, testratio, response, removelist, sampling)\u001b[0m\n\u001b[0;32m     80\u001b[0m                 \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     81\u001b[0m                     \u001b[0miter_x\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mjth_x\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 82\u001b[1;33m                 \u001b[0mregressor_OLS\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0msm\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mLogit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0my_train\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0miter_x\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdisp\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[0m\u001b[0;32m     83\u001b[0m                 \u001b[0mpval_list\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mj\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mmax\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mregressor_OLS\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mpvalues\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     84\u001b[0m                 \u001b[1;31m# Special condition where all the features have p values of 0, directly use these variables for training\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\statsmodels\\discrete\\discrete_model.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, start_params, method, maxiter, full_output, disp, callback, **kwargs)\u001b[0m\n\u001b[0;32m   1900\u001b[0m         bnryfit = super(Logit, self).fit(start_params=start_params,\n\u001b[0;32m   1901\u001b[0m                 \u001b[0mmethod\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mmethod\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmaxiter\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mmaxiter\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfull_output\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mfull_output\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1902\u001b[1;33m                 disp=disp, callback=callback, **kwargs)\n\u001b[0m\u001b[0;32m   1903\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1904\u001b[0m         \u001b[0mdiscretefit\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mLogitResults\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mbnryfit\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\statsmodels\\discrete\\discrete_model.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, start_params, method, maxiter, full_output, disp, callback, **kwargs)\u001b[0m\n\u001b[0;32m    215\u001b[0m         mlefit = super(DiscreteModel, self).fit(start_params=start_params,\n\u001b[0;32m    216\u001b[0m                 \u001b[0mmethod\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mmethod\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mmaxiter\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mmaxiter\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfull_output\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mfull_output\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 217\u001b[1;33m                 disp=disp, callback=callback, **kwargs)\n\u001b[0m\u001b[0;32m    218\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    219\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mmlefit\u001b[0m \u001b[1;31m# up to subclasses to wrap results\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\statsmodels\\base\\model.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, start_params, method, maxiter, full_output, disp, fargs, callback, retall, skip_hessian, **kwargs)\u001b[0m\n\u001b[0;32m    475\u001b[0m             \u001b[0mHinv\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mcov_params_func\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mxopt\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mretvals\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    476\u001b[0m         \u001b[1;32melif\u001b[0m \u001b[0mmethod\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;34m'newton'\u001b[0m \u001b[1;32mand\u001b[0m \u001b[0mfull_output\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 477\u001b[1;33m             \u001b[0mHinv\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mlinalg\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minv\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m-\u001b[0m\u001b[0mretvals\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'Hessian'\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m/\u001b[0m \u001b[0mnobs\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    478\u001b[0m         \u001b[1;32melif\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[0mskip_hessian\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    479\u001b[0m             \u001b[0mH\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m-\u001b[0m\u001b[1;36m1\u001b[0m \u001b[1;33m*\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mhessian\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mxopt\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\numpy\\linalg\\linalg.py\u001b[0m in \u001b[0;36minv\u001b[1;34m(a)\u001b[0m\n\u001b[0;32m    549\u001b[0m     \u001b[0msignature\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;34m'D->D'\u001b[0m \u001b[1;32mif\u001b[0m \u001b[0misComplexType\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mt\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32melse\u001b[0m \u001b[1;34m'd->d'\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    550\u001b[0m     \u001b[0mextobj\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mget_linalg_error_extobj\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0m_raise_linalgerror_singular\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 551\u001b[1;33m     \u001b[0mainv\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0m_umath_linalg\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minv\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0ma\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0msignature\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0msignature\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mextobj\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mextobj\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    552\u001b[0m     \u001b[1;32mreturn\u001b[0m \u001b[0mwrap\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mainv\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mastype\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mresult_t\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcopy\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mFalse\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    553\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mD:\\Anaconda3\\lib\\site-packages\\numpy\\linalg\\linalg.py\u001b[0m in \u001b[0;36m_raise_linalgerror_singular\u001b[1;34m(err, flag)\u001b[0m\n\u001b[0;32m     95\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     96\u001b[0m \u001b[1;32mdef\u001b[0m \u001b[0m_raise_linalgerror_singular\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0merr\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mflag\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 97\u001b[1;33m     \u001b[1;32mraise\u001b[0m \u001b[0mLinAlgError\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"Singular matrix\"\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     98\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     99\u001b[0m \u001b[1;32mdef\u001b[0m \u001b[0m_raise_linalgerror_nonposdef\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0merr\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mflag\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mLinAlgError\u001b[0m: Singular matrix"
     ]
    }
   ],
   "source": [
    "finalcolstouse = forwardselection(df,0.1,'Class','Time','naive')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### After variables that are deem significant are selected based on forward selection of logistic regression, they can be sent into the model pipeline for model evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Use existing model pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>V9</th>\n",
       "      <th>V10</th>\n",
       "      <th>...</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>-1.359807</td>\n",
       "      <td>-0.072781</td>\n",
       "      <td>2.536347</td>\n",
       "      <td>1.378155</td>\n",
       "      <td>-0.338321</td>\n",
       "      <td>0.462388</td>\n",
       "      <td>0.239599</td>\n",
       "      <td>0.098698</td>\n",
       "      <td>0.363787</td>\n",
       "      <td>0.090794</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.018307</td>\n",
       "      <td>0.277838</td>\n",
       "      <td>-0.110474</td>\n",
       "      <td>0.066928</td>\n",
       "      <td>0.128539</td>\n",
       "      <td>-0.189115</td>\n",
       "      <td>0.133558</td>\n",
       "      <td>-0.021053</td>\n",
       "      <td>149.62</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1.191857</td>\n",
       "      <td>0.266151</td>\n",
       "      <td>0.166480</td>\n",
       "      <td>0.448154</td>\n",
       "      <td>0.060018</td>\n",
       "      <td>-0.082361</td>\n",
       "      <td>-0.078803</td>\n",
       "      <td>0.085102</td>\n",
       "      <td>-0.255425</td>\n",
       "      <td>-0.166974</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.225775</td>\n",
       "      <td>-0.638672</td>\n",
       "      <td>0.101288</td>\n",
       "      <td>-0.339846</td>\n",
       "      <td>0.167170</td>\n",
       "      <td>0.125895</td>\n",
       "      <td>-0.008983</td>\n",
       "      <td>0.014724</td>\n",
       "      <td>2.69</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>-1.358354</td>\n",
       "      <td>-1.340163</td>\n",
       "      <td>1.773209</td>\n",
       "      <td>0.379780</td>\n",
       "      <td>-0.503198</td>\n",
       "      <td>1.800499</td>\n",
       "      <td>0.791461</td>\n",
       "      <td>0.247676</td>\n",
       "      <td>-1.514654</td>\n",
       "      <td>0.207643</td>\n",
       "      <td>...</td>\n",
       "      <td>0.247998</td>\n",
       "      <td>0.771679</td>\n",
       "      <td>0.909412</td>\n",
       "      <td>-0.689281</td>\n",
       "      <td>-0.327642</td>\n",
       "      <td>-0.139097</td>\n",
       "      <td>-0.055353</td>\n",
       "      <td>-0.059752</td>\n",
       "      <td>378.66</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>-0.966272</td>\n",
       "      <td>-0.185226</td>\n",
       "      <td>1.792993</td>\n",
       "      <td>-0.863291</td>\n",
       "      <td>-0.010309</td>\n",
       "      <td>1.247203</td>\n",
       "      <td>0.237609</td>\n",
       "      <td>0.377436</td>\n",
       "      <td>-1.387024</td>\n",
       "      <td>-0.054952</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.108300</td>\n",
       "      <td>0.005274</td>\n",
       "      <td>-0.190321</td>\n",
       "      <td>-1.175575</td>\n",
       "      <td>0.647376</td>\n",
       "      <td>-0.221929</td>\n",
       "      <td>0.062723</td>\n",
       "      <td>0.061458</td>\n",
       "      <td>123.50</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>-1.158233</td>\n",
       "      <td>0.877737</td>\n",
       "      <td>1.548718</td>\n",
       "      <td>0.403034</td>\n",
       "      <td>-0.407193</td>\n",
       "      <td>0.095921</td>\n",
       "      <td>0.592941</td>\n",
       "      <td>-0.270533</td>\n",
       "      <td>0.817739</td>\n",
       "      <td>0.753074</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.009431</td>\n",
       "      <td>0.798278</td>\n",
       "      <td>-0.137458</td>\n",
       "      <td>0.141267</td>\n",
       "      <td>-0.206010</td>\n",
       "      <td>0.502292</td>\n",
       "      <td>0.219422</td>\n",
       "      <td>0.215153</td>\n",
       "      <td>69.99</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         V1        V2        V3        V4        V5        V6        V7  \\\n",
       "0 -1.359807 -0.072781  2.536347  1.378155 -0.338321  0.462388  0.239599   \n",
       "1  1.191857  0.266151  0.166480  0.448154  0.060018 -0.082361 -0.078803   \n",
       "2 -1.358354 -1.340163  1.773209  0.379780 -0.503198  1.800499  0.791461   \n",
       "3 -0.966272 -0.185226  1.792993 -0.863291 -0.010309  1.247203  0.237609   \n",
       "4 -1.158233  0.877737  1.548718  0.403034 -0.407193  0.095921  0.592941   \n",
       "\n",
       "         V8        V9       V10  ...       V21       V22       V23       V24  \\\n",
       "0  0.098698  0.363787  0.090794  ... -0.018307  0.277838 -0.110474  0.066928   \n",
       "1  0.085102 -0.255425 -0.166974  ... -0.225775 -0.638672  0.101288 -0.339846   \n",
       "2  0.247676 -1.514654  0.207643  ...  0.247998  0.771679  0.909412 -0.689281   \n",
       "3  0.377436 -1.387024 -0.054952  ... -0.108300  0.005274 -0.190321 -1.175575   \n",
       "4 -0.270533  0.817739  0.753074  ... -0.009431  0.798278 -0.137458  0.141267   \n",
       "\n",
       "        V25       V26       V27       V28  Amount  Class  \n",
       "0  0.128539 -0.189115  0.133558 -0.021053  149.62      0  \n",
       "1  0.167170  0.125895 -0.008983  0.014724    2.69      0  \n",
       "2 -0.327642 -0.139097 -0.055353 -0.059752  378.66      0  \n",
       "3  0.647376 -0.221929  0.062723  0.061458  123.50      0  \n",
       "4 -0.206010  0.502292  0.219422  0.215153   69.99      0  \n",
       "\n",
       "[5 rows x 30 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "### From previous example above, we removed V18 in backward logistic selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount', 'Class']\n"
     ]
    }
   ],
   "source": [
    "# Based on backward selection and naive sampling, V18 was removed\n",
    "finalcolstouse = list(df.columns.values)\n",
    "finalcolstouse.remove('V18')\n",
    "print(finalcolstouse)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Logistic Regression without and with feature V18"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount', 'Class']\n"
     ]
    }
   ],
   "source": [
    "# Based on backward selection and naive sampling, V18 was removed\n",
    "finalcolstouse = list(df.columns.values)\n",
    "finalcolstouse.remove('V18')\n",
    "print(finalcolstouse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Oversampling is complete!\n",
      "[[55674  1194]\n",
      " [   10    84]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9788631017169341\n",
      "Sensitivity: 0.8936170212765957\n",
      "Specificity: 0.9790040092846592\n",
      "Precision: 0.06572769953051644\n",
      "f1 score: 0.12244897959183673\n",
      "AUC value is: 0.9766388830273617\n",
      "Model classification metrics have finished calculating!\n",
      "{'accuracy': 0.9788631017169341, 'sensitivity': 0.8936170212765957, 'specificity': 0.9790040092846592, 'precision': 0.06572769953051644, 'f1': 0.12244897959183673, 'auc': 0.9766388830273617, 'pr_auc': 0.6822765339665648}\n",
      "Model fitting and results are complete!\n",
      "Accuracy with all fraud results is 100.0%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1440 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# finalcolstouse is list of features that are significant (p < 0.05) for logistic regression + response\n",
    "# modelpipeline.run_model(self, df, varlist, response, testratio, standardize, sampletype, modelname, text, CV)\n",
    "results = modelpipeline.run_model(df, finalcolstouse, 'Class', 0.2, True, 'naive', 'LogisticRegression', 'Logistic Regression with backward linear selection', False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Oversampling is complete!\n",
      "[[55698  1170]\n",
      " [   10    84]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9792844352375267\n",
      "Sensitivity: 0.8936170212765957\n",
      "Specificity: 0.9794260392487867\n",
      "Precision: 0.06698564593301436\n",
      "f1 score: 0.12462908011869436\n",
      "AUC value is: 0.9768139805656698\n",
      "Model classification metrics have finished calculating!\n",
      "{'accuracy': 0.9792844352375267, 'sensitivity': 0.8936170212765957, 'specificity': 0.9794260392487867, 'precision': 0.06698564593301436, 'f1': 0.12462908011869436, 'auc': 0.9768139805656698, 'pr_auc': 0.6800490451604873}\n",
      "Model fitting and results are complete!\n",
      "Accuracy with all fraud results is 100.0%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1440 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "finalcolstouse = list(df.columns.values)\n",
    "# modelpipeline.run_model(self, df, varlist, response, testratio, standardize, sampletype, modelname, text, CV)\n",
    "results = modelpipeline.run_model(df, finalcolstouse, 'Class', 0.2, True, 'naive', 'LogisticRegression', 'Logistic Regression with backward linear selection', False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random Forest Models without and with feature V18"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount', 'Class']\n"
     ]
    }
   ],
   "source": [
    "# Based on backward selection and naive sampling, V18 was removed\n",
    "finalcolstouse = list(df.columns.values)\n",
    "finalcolstouse.remove('V18')\n",
    "print(finalcolstouse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Oversampling is complete!\n",
      "[[56768   100]\n",
      " [   18    76]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgEAAAERCAYAAADi2HRnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdd1wUR//A8Q9VmtiwYS+hKoJBxUgUbLHFXmN5rMGWqI/GGmPBrrFijT1REsMTjS0qRtGosYTExIIasYAFUSkKSN/fH/7YeIJy6OFx8fv2da+XNzs3O7vs3H5vdmbXSFEUBSGEEEK8dYz1XQEhhBBC6IcEAUIIIcRbSoIAIYQQ4i0lQYAQQgjxlpIgQAghhHhLSRDwLyATPERB8SaORTnexb+NPo9prYOACxcuMH78eBo3boybmxuNGzdm/PjxXLt2Ld8qd/r0aVq3bk2NGjVo1aqVzspdtmwZjo6OpKen66zMF/nhhx9wdHTE0dGRS5cu5ZjnypUrap5bt27lqfzvv/+eWbNm5ZqvcePGjBkzJk9lPy+//h5Zxo8fr+6HZ18eHh60bt2agIAAMjIydL7eFzl16hSOjo6cOHHija3zWbdu3cpxf2S9GjRooJd65SQqKorBgwcTGRmpszKf3/+PHj1i/PjxnDx5Us3zqse1LtqDtnr37k2PHj3eyLpelTbfiY6OjixatAj459j8/vvv31QV9Sq/vgtyajfP7uc3wVSbTN9++y3+/v7Url2bTz75hNKlSxMZGcmmTZvo1KkTa9eu5d1339V55ebMmcOTJ09Yvnw5RYsW1Vm5nTp1on79+piaarX5OmFsbMxPP/2Ek5NTtmW7d+9+5XJXrlxJ7dq1c823ZMkSbGxsXnk9kH9/j2cVL16cZcuWaaTFxcWxd+9eli1bRkpKCqNHj86XdRdUgwYNwsfHJ1u6ubn5m6/MC5w4cYLDhw8zYcIEnZXp4uLCli1bcHR0BCAsLIzt27fTtm1bna1DvJpSpUqxZcsWKleurO+qGLSc2s2WLVuwt7d/Y3XI9Sx49uxZpk+fTvfu3fniiy80lrVs2ZLOnTszduxYDhw4gImJiU4rFxcXh4eHB40aNdJpufb29m90JwO8++677Nu3j1GjRmVbtmfPHlxcXLh48WK+rb9mzZqvXUZ+/T2eZWZmhqenZ7b0pk2bcvv2bYKCgt66IKBSpUo57pN/u8KFC7+V220IzM3N5W+TT970fs31csDatWspXLgwn332WbZltra2jB8/njZt2vDo0SM1fe/evXTq1AkPDw/ee+89Jk+eTGxsrLp82bJlNGvWjGPHjtGhQwdq1qxJ48aN2bhxI/BPV9Pt27fZvXs3jo6O/PDDD2rX+s2bNzXq0aNHD3r37q2+DwsLo3///tSpUwd3d3c++ugjjh8/rrH+57u+XrfOuWnZsiU3btzIdqI/e/YsUVFRtGjRIttnDh48SM+ePfHw8KBGjRq0aNGCr7/+Wl2etY927dql/lrKqueaNWuoV68e7733HlFRURrdn3PmzMHR0VFjnxw8eBBHR8cct+dFfw+AGzdu8Omnn+Lt7U2tWrXo1asXZ86cyfbZzZs38+GHH1KrVi02b96s1T57Xk49Gd9//z0dO3bE3d0dNzc32rVrx969e9XlWd14p0+f5uOPP8bd3Z169erxxRdf8OTJEzWfoih89dVXNGnSBDc3N/r168e9e/eyrU/b7f3pp58YPnw47u7u1K9fn4CAAJKSkpgyZQp169alXr16fP7556SkpLzSvnhedHQ0EydOxMfHBzc3Nzp27MjBgwc18jg6OrJixQq6deuGm5sbc+fOBZ52SY4ePZp69eqp23T27FmNz/766690794dDw8P3n33XQYMGMD58+eBp8dc1i+Z5s2bZ+vJgX+Or7/++ktNO3DgAI6Ojqxfv15Ne/jwIU5OTuzZs0ejC/aHH36gT58+APTr14/x48ern8nIyGDRokV4e3vj5uZGjx49uHDhQq77LCMjg9mzZ1O3bl08PT357LPPePDgQbZ6v6wNAiQmJjJr1iwaNWpErVq1aN++Pfv27Xvhek+fPk2tWrUYOXIkly5dUo+XLBcuXMDR0RF/f381LT09HU9PT9asWQM8Pc7Gjh2Lt7c3rq6u1K9fn7FjxxITE6N+pnfv3owdO5bRo0fj4eFBt27dAHj8+DGTJ0/Gy8uL2rVrM3XqVFJTU3PdX896/nLADz/8gIuLC+fPn+ejjz7Czc0Nb29vFi5cqHEJLzMzk7Vr19K8eXNq1KhB06ZNWbduXbZr4tq26++//56mTZvi4eHBgQMHcqzry47dLL///ju9e/fG3d2dOnXqMHr06Bzb/7O0aTdpaWkEBATQrFkz3NzcaNGiBVu3bgVe3G6evxygbdsODAxk6tSpan0GDRpERETES7cBcgkCFEXhl19+oX79+lhaWuaYp3HjxowaNYpixYoBsGLFCkaNGoWLiwuLFi1iyJAh7N+/nz59+mh86T548IBJkybRuXNnVq5ciaurK7Nnz+b48eNqV1PJkiXx9vZmy5YtWv/6TEhIoF+/fhQqVIj58+ezePFizMzM8PPze+H1Sl3UOTf169enePHiGo0dnl4KyFr2rJCQEIYNG8Y777zDsmXLWLx4MeXKlWPGjBn89ttvANn2UZa7d++yY8cO5s2bx5gxYyhTpoxG2aNGjaJatWrqifDhw4dMnjwZb29v/vOf/2Sr+4v+HlevXqVjx47cuHGDcePGMW/ePAD69u3Lr7/+qlHG/Pnz6d69O3Pnzs31b5menq6+UlNTiYqKYs2aNRw/fpz27dur+bZs2cLkyZNp1KgRK1asYM6cOZiamjJmzBhu376tUebIkSNxcnIiICCAjz76iO+++45Vq1apyxcuXMjChQtp2bIly5Yto1q1akyaNEmjjLxs7xdffEHFihUJCAjAy8uLZcuW0aVLF1JTU5k/fz6dO3fm+++/Z9OmTS/dF/D0i/PZfZL1yvLgwQM6d+7MyZMnGT58OIsWLcLe3p5hw4axfft2jbJWrFiBt7c3ixcvpnXr1sTGxtK9e3f++OMPxo4dy4IFCzAxMaFPnz7qF2VERASDBw+matWqLFu2jNmzZ/Pw4UP69+9PQkICnTp1YvDgwQAsWrSITp06ZduGBg0aUKhQIY1rqln/P3XqlJr2yy+/YGJiwvvvv6/x+UaNGvH5558DMGHCBHV9APv37+ePP/5g+vTpzJgxg8jISAYPHpzr+JEDBw4QGhrKjBkzGDNmDEePHmXo0KHqcm3aYGZmJoMGDSIoKIg+ffoQEBCAi4sLI0eO5MiRI9nWefbsWfz8/PD19WXBggU4OTlRtmzZXPfLH3/8wePHj/H19SU5OZk+ffpw5coVJk6cyJo1a+jZsye7d+/myy+/1Fjf3r17SUlJYdmyZfj5+aEoCoMGDWL//v188sknzJs3j8jISDZs2PDSfaWNzMxMhg0bho+PDytXrqRx48asXr2a//3vf2oef39/Fi5cSPPmzVm+fDmtW7dmwYIFzJ8/X82Tl3Y9b948Pv30U6ZMmUKdOnWy1Sm3YxcgNDRUDTDnz5/PxIkT+eOPP+jVq5ea53natBt4Os5p1apVfPjhhyxfvpymTZsybdo0AgMDtWo3eWnbixYtIi4ujrlz5/L5559z9uxZxo4d+9K/GQDKSzx8+FBxcHBQ5s+f/7Jsqri4OKVGjRrK+PHjNdJPnTqlODg4KJs2bVIURVGWLl2qODg4KCEhIWqepKQkxdXVVZk+fbqa5uvrq4wePVp9/7///U9xcHBQbty4oVF+9+7dlV69eimKoihnz55VHBwclNDQUI3tmDVrlnLlyhWN9aelpem8zs97ts5TpkxRmjZtqi5LT09XGjRooGzfvl3Ztm2b4uDgoERGRiqKoiirV69WRo0apVFWTEyM4uDgoCxfvvyF+yineuaU76+//lJcXFyUBQsWKMOGDVPq1aunREdHv3A7cipjxIgRSp06dZT4+Hg1LTU1VWnevLnSvn17RVEUJTIyUnFwcFBGjBjx0rIVRVHGjRunODg45Pjy9fVVVqxYoaSmpqr5Z86cqcycOVOjjHPnzikODg7Kjh07FEVRlJMnTyoODg7K3LlzNfL16NFDadOmjaIoivL48WPF1dVV8ff318gzceJExcHBQTl+/Hiet3fYsGFqnnv37ikODg5Kly5dNMpv1aqV4ufn98L9kVXWi17JycmKoijKvHnzFFdX12ztonfv3oqXl5eSlpamKIqiODg4KJ07d9bIs3DhQsXV1VW5fv26xja1atVK6d+/v6IoirJnzx7FwcFBiYqKUvNcvXpVmTt3rnL37l1FUV7cNp81YMAAtZ0qiqI0b95c6dixo1K7dm0lPT1dURRFGTVqlNK7d29FUf7522Xt/+ffK8rTY7JBgwbqvlAURdmyZYvi4OCgXLp06YV18fX1VerWras8fvxYTQsODtZoO9q0wZCQEMXBwUHZs2ePRr7//Oc/yowZMxRFUZRevXop3bt3Vy5cuKB4enoqn376qfo3URRF+eKLLxRfX1/1ff/+/ZWOHTsqjo6OysOHDxVFUZQFCxYojRs3VhRFUS5evKh069ZNCQ8P11inn5+fxvdLr169FBcXF41tPHr0qOLg4KAcOHBATUtPT1c++OAD9TvxRRwcHJSFCxcqivLPsblt2zZFUf75+3/zzTdq/szMTKVRo0bqMX79+nXF0dFRWbZsmUa5K1euVJydnZU7d+4oipK3dj1v3rwX1ldRtDt2u3fvrjRv3lxJSUlR80RERCiurq7KqlWrNNaXdexp026uXr2qODg4KGvWrNGo07hx45Thw4dr7Ldn282z+zkvbbtTp04aeRYtWqQ4ODgoMTExL91HL+0JyLrGr+2I7LNnz5KamsqHH36okV63bl3KlSvH6dOnNdKfHUxoaWlJ0aJFSUpK0mpdL/LOO+9QsmRJhgwZwrRp0/j5558xNzdnwoQJvPPOO3qtc6tWrYiIiFAjxVOnTvH48WOaNm2aLe/HH3/MwoULSU5O5tKlS+zbt4+1a9cCaNV1l3V54EVq1qyJn58fa9euJTg4mJkzZ1KyZEmttiPL6dOnadiwIba2tmqamZkZbdq04eLFizx+/FhNd3Bw0KrMEiVKEBQURFBQEBs3bsTb25siRYrwxRdfMGTIEMzMzNS8EydOZOLEiSQkJHDu3Dl2795NYGAgkH0fPT9wtWzZsurf7Y8//iAtLY1mzZpp5GnTps0rb++z6ytVqhQA7u7uGuUVK1aM+Pj4XPfJkCFD1H3y7CtrYODp06epWbMmlSpV0vhc+/btiYmJ4erVq2ra83+HkydP4uDgQPny5dUeBiMjI3x9fTl16hSpqam4u7tjZWVF165dmTt3LsePH6d8+fKMHTs2Wy/Ty/j6+vLHH3/w5MkT7t69y40bNxgyZAgJCQlcuHCBjIwMjh8/TuPGjbUuE6BWrVoUKlRIfV+xYkUAjUuUOfHx8dG4xNS4cWNMTU3VX+LatMHffvsNY2PjbMfOxo0bNXqS7t27R//+/UlOTmbatGkag5J9fX25ffs2ERERpKamEhoayqBBgzA2NlZ7A44cOaLuF2dnZ7799luqVKlCZGQkx44dY8OGDVy7di3bcW9vb6+xjadPn8bExARfX181zcTEhJYtW750X2nr2ePeyMiIMmXKqO3s5MmTKIpC06ZNNXq0mjVrRkZGhjrrIy/tOrfvldyO3eTkZM6ePYuvry/GxsZqncqWLYurqyvHjh3LsVxt2k3WZcLnL/XOmTMnx0tmOclL287pOw7Q6M3OyUsHBhYpUgRra+tsXTDPSklJ4dGjR5QsWVL9QrOzs8uWz87OLlujtLCw0HhvbGz82vMlraysCAwMZPXq1ezfv5+tW7dibm5O8+bNmTp1KoULF9bI/ybr7OnpSenSpfnpp5+oUaMGu3fvzvZFlCU2NpZp06YRHByMoigag8O0WV9O2/O8Dh06sGLFCooUKZJjV1pu4uPjcwwcstb97ElR2wDD1NRUYxCjp6cnffr0Yfjw4WzatEnjQI+MjGTq1KkcP34cU1NTqlWr9sLg5/nLWUZGRup+jIuLA8h2SSbr5J0lL9trbW2tVR20Ua5cuZcO7IyPj89x1klO9Xr+uIiNjeXmzZu4urrmWHZsbCz29vZs3bqVr776iqCgINavX4+VlRUdO3Zk3LhxWs9S8PX1Zfr06Zw5c4b79+9TqlQpmjRpgp2dHadPnyY9PZ24uLg8BwHP71dj46e/bTIzM1/6uef3hbGxMcWKFVPbvDZtMDY2liJFimgEpzm5ffs2DRo04MyZMyxcuJDp06ery7Iutx4/fpwqVaqQnp5Ow4YNcXFx4fTp09SuXZvLly9rjCDftGkTq1ev5uHDh9jZ2VGjRg0sLS1JTk5+6TbGxcVha2ubbWbU88f6q3pZO8saY9WuXbscP5t1DT4v7Tq375Xcjt34+HgyMzPZsGFDjpdEXjT7QZt2k7W9JUqUeGkdXyYvbTuncxPk3g5ynR3g7e3NqVOnSE5OzrYSgJ9++olx48axZs0aihQpAjy9jvF8hBYdHU2tWrVyW91LZX1pPt8zkZCQoDFlrUKFCsyYMQNFUbh06RJ79+5l3bp16i/KZ+V3nZ9lbGxMixYt2LdvHyNGjCA4OPiFc/xHjx5NeHg4GzZswN3dHXNzc548ecK2bdt0UhdFUfjiiy8oW7YsiYmJTJ8+nQULFuSpjCJFinD//v1s6dHR0QAULVpUY6DSqzAzM2POnDl8+OGHjBs3jj179lCoUCH1WqyZmRlBQUE4OTlhamrK1atX+fHHH/O0jqyT/4MHDzR6i54dGAp5215tT/C6kFu9ssbr5KRw4cK8++67L5zal/VZZ2dndZDXX3/9xfbt2/nmm2+wt7dnwIABWtXT3t4eBwcHTpw4QUxMDPXq1cPIyIi6dety+vRpEhISqFatmvpLPr893wuTnp5ObGysejxo0wYLFy7Mo0ePSE9P1zixXr58mSdPnqi9P1WrVmX16tWsXr2agIAA2rRpQ926dQEoVKgQXl5enDhxgujoaGrWrImVlRVeXl4cPnwYFxcXjZkSu3btYtasWXz22Wd07NhRre+IESOyHbPPK168OPHx8aSlpWkELrl9TheyetA2bNiQ7ccYPA1EdNmus7zs2O3evTtGRkb07t07x6mnLwpwtWk3WdsbExODlZWVuiwiIoKoqCitZgG8TtvWVq6zA/r37098fHyOJ4i4uDiWL19OmTJlaNCgAbVq1cLc3Jxdu3Zp5Dtz5gx379597XsJZP1ivnv3rpp2//59rl+/rr7ft28fXl5eREdHY2RkhLOzM6NHj6Zq1ao59mjkd52f16pVK27dusWaNWtQFOWFg+RCQ0Np2rQpdevWVQ/ErIFGz/YEZEV7ebVlyxZOnDiBv78/n332Gbt27WL//v15KqNOnTocPXpUo7ckPT2dvXv34uLionHgv45KlSrx8ccfExkZqY6Ojo2N5fr163To0IEaNWqoX8BZ+yi36PdZHh4eWFpasmfPHo30n3/+WeP9m9revKpTpw7nzp3LNhJ4586dFC9enCpVqrzws3Xr1uX69etUqlSJmjVrqq+9e/eyceNGzMzM2LhxI76+vqSkpGBiYoKHhwfTpk3DyspKbVPaHoc+Pj6cOHGC0NBQvLy8APDy8uK3337j8OHDL+0F0PUU5GPHjml0L+/fv5/09HTq1asHaNcGPT09ycjI4NChQxplT5s2TWOQXtGiRTEzM+Pjjz+mcuXKTJ48WeNXu4+PD6dOneLMmTMa++Xq1av88MMPvP/+++pJOzQ0FCsrKwYOHKgGAAkJCYSGhuZ63L/33ntkZmZmG6D8/LGeH7J6Gx8+fKhxrCUlJbFgwQKio6N12q6BXI9da2trXF1dCQ8P16iTo6Mjy5cvz3FwJ2jXbrLOHcHBwRqfXbp0KRMmTMDY2DjXdvM6bVtbufYEuLu7M3LkSBYtWsS1a9do164ddnZ2XL16lU2bNhEbG8vmzZsxNTWlaNGifPzxxwQEBGBmZkaTJk24desWS5cupUqVKjmOfsyL+vXrY2VlxZw5cxgxYgSpqamsXr1aI6qsXbs2iqIwePBgBg4cSJEiRfjll1/4+++/GThwYLYy87vOz3N3d6dcuXLqiNEXRZpubm7s3r0bV1dXypQpw++//87atWsxMjLSGINga2vL5cuX+fXXX9Uvr9zcvHmTBQsW0KFDB7y9vYGnvy6mTJmCp6en1t1Xw4cP5+jRo/Tq1YuPP/6YQoUK8fXXXxMZGcnq1au1KkNbgwYNYseOHXz11Ve0b9+eChUqUK5cObZu3UqZMmWwtbXl2LFjfPPNN0Du18GeZWVlpY6UtrCw4P333yc0NJTvvvtOb9ubF/369WPnzp307duXoUOHUrx4cXbs2MGpU6eYMWPGS0+eWZ/t06cP/fr1o0SJEhw8eJDAwEBGjBiBkZERXl5efPnllwwePJiePXtibm7O7t27SU5OVq93Zv3qCQ4Opnnz5i/8Ne/j46MGclnHq5eXF4mJiVy6dClbT92zstr50aNHKVmyZI5jfPIiLi6OIUOG0KdPHyIiIli0aBFeXl7qzARt2qCPjw8eHh5MmjSJe/fuUalSJfbt28fZs2dZt25dtnWam5szbdo0+vTpw9KlS9XR276+vkyZMoUzZ84wfPhw4Ok1XjMzM37//XeN0fO1atUiMDCQGTNm0LhxY6Kjo1m/fj0PHjzI8TLUs+rVq4ePjw9Tpkzh4cOHVKlShR9++IHw8PDX2pfacHBwoH379nzxxRdERkZSq1YtIiIiWLJkCXZ2djg6OmJubq6zdg1odeyOHj2agQMH8umnn6q9AV9//TW//fYbffv2zbFcbdqNk5MTLVq0YNGiRSQnJ1OjRg1OnTrFrl27mD17NpB7u3mdtq0trcL3wYMHs3btWszNzdUd+s033+Dl5cXOnTs1rld+8sknTJ06ldDQUIYNG8bKlStp0aIF3377ba4HaG5sbGxYvnw5ZmZmjBw5koULF9KtWzeNXw+lSpVi/fr1lChRgunTpzN48GBOnDjBrFmzNKaXPSs/65yTVq1akZaWlm0w4rPmzJmDh4cHs2bN4tNPPyUkJISpU6fSsGFDQkND1Xx+fn7cv3+foUOHcufOnVzXnZmZybhx47C2ttaYaz1jxgyePHnC5MmTtd6Od955h61bt1K6dGkmT57M2LFjMTIyYtOmTdmmeL2uQoUKMXHiRFJSUtQGtGLFCsqWLcukSZMYPXo0586dY/ny5VSvXl2dwqWtAQMGMG3aNH755ReGDx/OqVOnmDlzpkaeN7m9eVGyZEkCAwNxc3Nj3rx5jBo1iqioKAICAujSpctLP1uqVCl1kNmsWbMYNmwYoaGhTJkyRZ0u5+TkxJo1a0hLS2PChAl88skn3Lhxg+XLl6td2vXr11enHuZ08svi4eFBsWLFKFeuHBUqVACe9vTY29tTrFgxPDw8XvjZd955h3bt2rFlyxZ1eubr6NatG+XLl2fUqFEEBATw4YcfsmLFCvVSjjZt0MTEhLVr19KmTRtWrVrFp59+yt9//82qVauoX79+juutV68eHTp0YOPGjZw7dw6A0qVL4+Ligrm5uboPLC0tcXd3x8TEhIYNG6qfb9++PcOGDSM4OJjBgwerfwd/f38SEhJeeHvyLEuXLqVr166sXbuWkSNHYmxszJAhQ157f2pj1qxZ+Pn5sWPHDvz8/FixYgXNmjVj48aN6g8iXbZrbY7d9957jw0bNhAbG8uYMWMYP368ej+DrF6Z52nTbuDplMN+/frx3XffMXToUI4cOcK8efPo2LEjkHu7eZ22rS0j5XVH4gkhhBDCIMlTBIUQQoi3lAQBQgghxFtKggAhhBDiLSVBgBBCCPGWkiBACCGEeEvlep8AIYQQIq+MmpXXOq8SfCsfayJeRoKAt0xeGqb498v68k3OeL0Hd4l/FwsTHdz98g3eOlu8OgkChBBC6J5cbDYIEgQIIYTQPekJMAgSBAghhNA9iQEMggQBQgghdM9EogBDIEGAEEII3ZPLAQZBggAhhBC6JzGAQZAgQAghhO4ZSxRgCCQIEEIIoXsSAxgECQKEEELonowJMAgSBAghhNA9mR1gECQIEEIIoXsSAxgECQKEEELonlwOMAgSBAghhNA9mR1gECQIEEIIoXsSAxgECQKEEELonvQEGAQJAoQQQuieBAEGQYIAIYQQuicxgEGQIEAIIYTuyewAgyBBgBBCCN0z1ncFhDYkCBBCCKF7+dQTcOfOHUaOHJkt3c/PjyZNmnDjxg02btxIeHg4NjY2tGzZkrZt26r5MjMzCQoK4ueffyYxMREnJycGDhxImTJl1Dy6KMNQSBAghBBC9/JpYODNmzextLRkyZIlGulWVlY8evQIf39/6tSpw8CBA7l69Spr167FysqKpk2bAhAUFMSBAwcYOnQoJUqUYOvWrcycOZMvv/wSc3NznZRhSKTDRgghhO4Z5+GVBxEREZQvX56iRYtqvMzNzTl48CCmpqYMGjSI8uXL4+PjQ5s2bfjxxx8BSEtLY/fu3XTp0oXatWtTqVIlRowYQVxcHCdPngTQSRmGRIIAIYQQumdkpP0rD27evEm5cuVyXHbp0iWcnJwwMTFR01xdXbl37x4xMTHcuHGD5ORkXF1d1eVWVlZUqVKFixcv6qwMQyKXA4QQQuheHs7tiYmJJCYmZku3trbG2tpaIy0yMpIyZcowefJkoqKiKFu2LJ06daJWrVrExMTg5uamkb9YsWIAPHjwgNjYWABKlCiRLc/Dhw8BdFKGIZEgQAghhO7lYUzAnj17CAoKypbeuXNnunbtqr5PTk4mOjoaW1tbPvroIywtLTl69CizZs1i0qRJpKSkYGqqeVozMzMDnnbjp6SkaKQ9myctLQ1AJ2UYEgkChBBC6F4euvlbt26Nj49PtvTnewEsLCzYtGkTpqam6om6atWq3L59m507d2Jubk56errGZ7JOzBYWFuqgvbS0NI0TfVpaGhYWFgA6KcOQSBAghBBC54zy0BOQU7f/i+R0oq1YsSKhoZNawg0AACAASURBVKGULl2amJgYjWXPdt8riqKmWVpaauQpX748AHZ2dq9dhiGRgYFCCCF0zsjISOuXtq5cuUKfPn24evWqRnp4eDjly5fH2dmZy5cvk5GRoS67cOECZcuWpWjRolSqVAlLS0suXLigLk9KSuL69eu4uLgA6KQMQyJBgBBCCJ3Lj8kBVatWpVSpUqxevZpLly5x+/ZtNmzYwOXLl+nUqRO+vr4kJyezcuVKbt26xdGjR9m9ezcdOnQAnl63b9GiBYGBgZw5c4abN2+yZMkSihcvjpeXF4BOyjAkRkpW34Z4Kxg1M7zuKpF/lOBbACRnJOm5JqIgsTCxeu0yzP/rrnXe1IVntc4bExPD1q1b+fPPP0lKSqJq1ar06NFD/RUeHh7Ohg0buH79OkWLFqV169a0atVK/XxmZiaBgYGEhISQnJyMs7MzAwYMoHTp0moeXZRhKCQIeMtIECCeJUGAyIkugoBCoz20zpvy5R+vvT7xamRgoBBCCJ0zNparzYZAggAhhBA6J08SNgwSBAghhNC5vIz6F/ojQYAQQgidkyDAMEgQIIQQQueM8vLwAKE3EgQIIYTQOekJMAwSBAghhNA5kzzcNljojwQBQgghdE56AgyDBAFCCCF0ToIAwyBBgBBCCJ2TGMAwSBAghBBC56QnwDBIECCEEELnJAgwDBIECCGE0Dl5doBhkCBACCGEzklHgGGQUE0UeMULF0UJvpXt9f3k1Wqebj5t+WvNQZ7sucqVjb8wvF0/ddmU3v/N8fNK8C2ubT6h5jMxNmFqn9Hc3HKKxF1/c2rZbhp7NNCoi3PFd/hp1jfEbr9AxJbTLBk6HWuL13/sqsh/IYdCqO+p+fdUFIWvVq3lg8YtqVe7Pn4DBnP92nWNPKmpqcyfs4DG7zelvmcDRo8cQ3R09JusukEyMjLS+iX0R3oCRIFXq5oLAM3Hf8SjpAQ1/eGjWAC6NvqQrRMC+DJoDSNWHKKJhzfLhvvzKOkxm4ODWPtTIPt+C9Eo06lCNdaP/pK1+wLVtKXDptO7aSfGfjWT8Ls3GdKmD7v9N+Ex5AMuR4ZT1KYIB+cFcufhPXrO/oSSRUswf9Dn2JcoTRd/v/zfEeKVnf3jLBPHfY6iKBrpq1esYf3aDYz476eUK2fPmtVrGdTfj+27/kfhwoUB8J86kyOHjzB67H+xsrJk6aJlDPf7hMCgrZiYmOhjcwyCnNwNgwQBosBzq+JMVEw0waFHc1w+b9AkVuzazNivZgBw+OwJKpeuQLPaDdkcHMTtB3e5/eCumt/Y2JiA4f4c+esks7YuA6B6uSoMbtObrjOG8L9f9gAQ8uev/Ln6AE08vLkcGU7zdxtiX6IM3iM7cj0qAgBzU3PWjJpLCdtialAiCo7U1FS2fL2V5UtXYGlpSWZmprosMTGRTRs2M3iYHz17fwRA7Xdr06JpK7b/bwd9+vYmMiKS3Tt3M3v+LFq0/AAAB0cH2rXuwOFDITRt1kQv22UIjCUIMAgSBIgCz62qM39dD8tx2bsOblQqXZ41e7ZopPea88kLyxvYsge1qrrgPvgDNa1d/ebEPI5TAwCAtPQ0XAb4qu8LmZkD8CjpsZqWdeIvXrioBAEF0LGjx1n31XpGjRlJfFw8mzd+rS77689zJCUl4ePbSE2zLWLLu3Xe5fixE/Tp25vTp84A0LBRQzVPpcqVqFa9GieOnZAg4CWM5bbBBkGCgALk0aNHHDx4kLCwMB48eEBaWhoWFhaUKFECFxcXmjRpgo2Njb6r+ca5VXUmOTWF44t3UPudGjyIj2XpjvXM37YStyrOAJiamBDyZRD1nWtzL/YBs78NYOWuzdnKKmRWiKm9/8v6fd9x8eYVjXVcirxKp/db4993DO+Uq8L5G5cZsWIKR/86CcCukweJiL7NoiFTGb16OsULF+WLXiP57cqf/H37erZ1Cf1zrenK3gN7sLUtzMqAVRrLbt64CUCFChU00suXL0fIoRA1Twm7ElhZWWbLk/V5kTN5iqBhkCCggAgPD2fmzJlYWlri7OxM5cqVMTMzIy0tjdjYWA4cOMCuXbv4/PPPqVy5sr6r+8YYGRnhUtGBxOQkxqzxJyL6Dq3qNmZ2//FYmBciJS2V9Ix0dk7fwIpdm5j29UI6NGjJik9n8fBRLNuO7NIor7tvW0oVtWNB0GqN9JJFSvBOuSosHjKVievncC/uAWM6+/HTzK9xGejLzXu3iEuIx2/xeP73xRp6N+0EwPW7Ebz/305vbH+IvCldutQLlyUmJGBubo6ZuZlGurW1NQkJiQAkJCRibW2d7bNW1tZERd3TbWX/ZWRMgGGQIKCA2LBhA56engwePDjH+bWZmZmsWrWKDRs2MG3aND3UUD+MjIxoM/k/RETfIfzODQBC/jyBjaUV47oOZfa3AZiamLJm7xZmBwYAT8cEVClTgSm9R2ULAj5u1ZOfzhzm6nO/3M1MTSldrCQN/9uJX86dAuDY+dOEbzrO2K5DGLZsEh94+rBz+np2nTzIqt1fY2NpzeSeIwieuxXvkR2IeRyX/ztE6Iyi5HyiUhTlmTao5HwyUxTp7s6FBAGGQaYIFhA3btygXbt2L7zBhrGxMW3btuXatWtvuGb6lZmZyeGzJ9QAIMu+MyFYW1ppvH9W8O+/4FC+Kmam//zKK1XUDi/n2nwbsjPbehKeJJL4JEkNAACSkp/w68XfqVnFCYAxXfwIi7hKF38/gkOPsv3YTzQf/xHl7coyouNAHWyteJNsCtuQmppKWlqaRnpSUhI2hZ9edrOxsSExMTHbZ5OSkrCxKfxG6mmojIy0fwn9kSCggChWrBg3b778GuPVq1fVaUtvi7IlSjOoVU/sihTXSLcsZAHAjXu3ADA31ezSNTMxxQgjjdHgH3g2IiMzg52/Hsi2nqt3bmBiYoyJseaULzNTU3VaWYWS9py5/KdGmQ/iY7h8KxyXiu+8xlYKfahYqSKKonD79h2N9Fu3blO5ciU1z8MHD0lOTs6ep0qlN1ZXQyT3CTAMEgQUEK1bt2blypVs3bqVs2fPEhkZSVRUFJGRkfz1118EBgaybt06WrVqpe+qvlGFzJ5OwevVpKNGeqf3W3E5Mpydvx7gSUoyXRq10Vjeul4Tzlz5k4zMDDWtrpMHYRFXefzMvQayHAg9ioW5BR/Wb6amFbG25T0XT05cDAXgyq1r1HNy1+itKVa4KO+Uq8L1qEidbK94c9zda1GoUCEO/3xYTXsU/4jQM6HU9aoLQD2vumRkZHDk8D/TU2/euEn41XDq/X8ekTNjY2OtX0J/ZExAAdGiRQssLCzYvn07P/74Y7bl9vb29O3blyZN3q4pSTeiItl6aAf+fT8jU1EIi/ibLg3b0Mm7Fe2nDuBxUgKzApcxtfd/eZT4mCN/naSbT1sauXnRalIfjbJqVHbk8q3wHNcTHHqUg7//wrr/LqCEbTHuPLzHhO7DAFiyfR0AswKXceTLIP73xRpW7v4aawsrJnQfRkZGJst2rM/fHSF0zsraih49uxOwZDnGRsZUqlyRr1avw9rGmo6dOwBQoWIFmn/QjOlT/ElISMDWtjBLFy3DweEdfJv45rKGt5v8wDcMRsrzt9ASehcfH09MTAwpKSmYm5tjZ2eHra2tTso2alZeJ+W8SRbmFkzuNYIePu0oW6IUYRFXmf7NYnYc36fm+aR9fz5p348KJcty5dZ1pmz+UmM5wIW1hzh+4Tc+XjQ2x/VYW1gxe8AEujZqg42FNb+GhTJy5VQu3Lis5mno5sWMvp/h6VCL+MRHHDt/hrFfzVRvHmRolOCnl1OSM5L0XJP8tzJgFZs2bOZk6D+3ik5PTydg6XJ2bt9FUlIStTxqMX7iWKpUraLmSUp6woK5CwjefxBFyaRe/XqMmziWUqVePPPA0FmYvP6tsJ0Wt9Q676WRP732+sSrkSDgLWOIQYDIP29TECC0p4sgwHmJ9pcuw0bsfaV13Llzh3Hjxmn0kkZHR7N+/XrCwsIwNzenUaNG9OjRQ+MWz/v27WP37t3ExsZSuXJl+vXrR/Xq1dXluijDUMjFGCGEEDqX3wMD09PTWbZsGSkpKRppM2fOBMDf3x8/Pz8OHz7Mtm3b1DwhISFs2bKF7t27M3fuXOzt7Zk5cybx8fE6K8OQSBAghBBC5/J7iuC2bduwsLDQSDt58iT3799n+PDhVKxYEU9PT3r27MnevXtJTU0FYPv27XzwwQd4e3tTvnx5hgwZgpWVFcHBwTorw5BIECCEEELn8nN2wMWLFzl48CDDhg3TSA8LC6Ny5coat1d3cXEhJSWFa9euERcXx927d3F1ddWop5OTE2FhYTorw5DI7AAhhBA6l5du/sTExBxvymRtbZ3tts2JiYkEBATQv39/7OzsNJbFxMRQvLjmPUWy3j98+BBz86cPAStRooRGnmLFihEeHq6zMgyJBAEFxJQpU7TO+zbdNlgIYZjy0s2/Z88egoKCsqV37tyZrl27aqR99dVXODg44O3tnS1/ampqtplUpqZPT3NpaWlqd35WWpas57ToqgxDIkFAAVG6dGl9V0EIIXQmLz0BrVu3xsfHJ1v6870AR48e5dKlSyxYsCDHcnI6EaenpwNgYWGh/orPSsuS9cRWXZVhSCQIKCCGDh2q7yoIIYTu5CEIyKnbPyeHDx8mPj6eIUOGaKSvX7+ePXv24OLiwvXrmg8Hi4mJAZ526WddPoiJiaFixYpqntjYWLXL387O7rXLMCQSBBRQcXFx3LlzR71PvaIopKWlcfXq1WzdY0IIUdDkx1MWP/nkE7U7Psunn35K586d8fb25sqVKxw+fJjExH8eAX3hwgUsLS2pWrUqpqam2Nvbc/HiRdzd3YGnDym7dOkSTZs2BcDZ2fm1yzAkEgQUQEeOHGHNmjXZupsASpUqJUGAEKLAy48HA73ol7atrS0lS5akSJEiFC9enMWLF9OrVy/u37/P1q1bad26tXoNv02bNmzYsAF7e3uqV6/Ozp07SU5OVm82VKdOndcuw5BIEFAA7dixgwYNGtChQwcmTpzI5MmTiYmJYd26dXTp0kXf1RNCiFzp4+mA5ubmTJo0iXXr1jFx4kRsbGxo1qwZnTt3VvM0bdqUJ0+e8N133/H48WOqVq3KpEmT1MGAuijDkMhtgwugnj17Mm/ePMqVK8e0adNo164d7u7unDp1iu3btzNnzpxXLltuGyyeJbcNFjnRxW2D66zvnHum/3emf/aZAeLNkJsFFUCmpqaYmZkBT2cNREQ8fThNtWrVuHv3rj6rJoQQWsnvOwYK3ZAgoACqVq0aBw4cAKBixYr8+eefAERERGSbmyqEEAVRfj87QOiGnFEKoC5dujBr1ixsbW3x8fHh+++/Z8SIEcTGxvLee+/pu3pCCJGrV7kdsHjzJAgogJydnVmyZAnp6enY2tri7+9PcHAwJUqUoFUr7R/PKYQQ+iK/8A2DBAEF1LNTYcqXL0+/fv30WBshhMgbiQEMgwQBBVBuzwbIy3MGhBBCH6QnwDBIEFAAPX9DjIyMDKKiorh9+zYtW7bUU62EEEJ7EgQYBgkCCqBPPvkkx/Rvv/2WhISEN1wbIYTIOwkCDIMM3zQgjRs35vjx4/quhhBC5MrY2Ejrl9Af6QkwIJGRkfqughBCaEd6AgyCBAEFUEBAQLa0pKQkzp8/j5eXlx5qJIQQeSOXAwyDBAEF0IMHD7I1IFNTU9q2bUubNm30VCshhNCe9PIbBgkCCqCpU6e+cFlcXBwWFhZvrjJCCPEKpCfAMMjAwAKoW7duPHr0KFv6gwcPXjhzQAghChITY2OtX0J/pCeggDhx4gRnz55V369fvx5zc3ONPPfv35deACGEQZBTu2GQIKCAcHBwYP/+/SiKAjw94ZuYmKjLjYyMsLCwYNiwYfqqohBCaM1YLgcYBAkCCgg7Ozv1dsHTpk1jzJgxWFtb67lWQgjxamRMgGGQHpsCaMqUKYSFhREaGqqmrVq1ij/++EOPtRJCCO0ZGxlp/RL6I0FAAXTo0CEWLlzI/fv31bT09HTmz5/PiRMn9FgzIYTQjpGRkdYvoT9yOaAA2rlzJwMGDKBJkyZq2vDhw3FycuKHH37gvffe02PthBAid6ZycjcI0hNQAD148ICaNWtmS3dzcyMqKkoPNRJCiLyRngDDIEFAAVSyZEnOnz+fLT0sLIxixYrpoUZCCJE3MibAMMjlgALogw8+YMOGDURFRVG9enUAwsPD+emnn+jSpYueayeEELmTU7thkCCgAGrRogWpqans2bOHH3/8EYBixYrRs2dP6tWrp+faCSFE7uQXvmGQIKCAatu2LW3btuXx48eYmppy7do1Dhw4wObNm9myZYu+qyeEEC8ltwM2DBIEFGBJSUn88ssvBAcHc+fOHczMzGjUqJG+qyWEELmSngDDIEFAARQeHk5wcDDHjx8nNTUVgFatWtGhQwdsbW31XDshhMhdfoUAMTExbN68mXPnzpGamoqLiwu9evWiQoUKANy4cYONGzcSHh6OjY0NLVu2pG3bturnMzMzCQoK4ueffyYxMREnJycGDhxImTJl1Dy6KMNQSH9NAZGamsqhQ4eYMGECEydO5Ndff6VevXqMGzcOY2NjmjRpIgGAEMJg5MfsAEVRmD17NjExMUyaNInZs2djbm7O9OnTSU5O5tGjR/j7+1OmTBlmz55Nt27d2LZtGwcPHlTLCAoK4sCBA/j5+TFz5kxMTEyYOXOm+oNLF2UYEukJKCD8/PxQFAV3d3fat29P7dq1MTMz03e1hBDileTH5YD4+HjKlStH165dsbe3B6BTp06MHTuWiIgIzp8/j6mpKYMGDcLExITy5csTFRXFjz/+SNOmTUlLS2P37t307NmT2rVrAzBixAj8/Pw4efIkDRs25ODBg69dhiGRnoACIjMzE0tLSwoVKkR6ejoZGRn6rpIQQryy/LhZUNGiRRk5cqQaAMTFxbFr1y6KFStGxYoVuXTpEk5OThpPYHV1deXevXvExMRw48YNkpOTcXV1VZdbWVlRpUoVLl68CKCTMgyJ9AQUEGvWrOHXX3/l8OHDhISEYG5ujoeHh9wiWAhhkEzycHJPTEwkMTExW7q1tfULn6a6fPlyjhw5gpmZGWPHjsXCwoKYmBjc3Nw08mXdYO3BgwfExsYCUKJEiWx5Hj58CKCTMgyJBAEFRKFChfDx8cHHx4e7d+9y6NAhjh49yqlTp4CnzxNo164d5cqV03NNhRAid3m5HLBnzx6CgoKypXfu3JmuXbvm+Jm2bdvSokUL9u/fz/z585k+fTopKSmYmmqe1rIuq6alpZGSkqKR9myetLQ0AJ2UYUgkCCiAypYtS8+ePenRowe///47ISEhHDt2jCNHjlCjRg0mT56s7yoKIcRL5SUIaN26NT4+PtnSX9QLAKizAQYPHszff//N3r17MTc3Jz09XSNf1onZwsICc3NzNe3ZE31aWhoWFhYAOinDkEgQUIAZGxvj6emJp6cnjx49IiQkhJCQEH1XSwghcpWXa/0v6/Z/VmxsLBcuXKBBgwZq+cbGxlSoUIHY2Fjs7OyIiYnJ9hl42n2vKIqaZmlpqZGnfPnyADopw5DIwEADYWtrS9u2bVm4cKG+qyKEELkyzsNLWw8ePGDp0qVcvnxZTUtPT+f69euUK1cOZ2dnLl++rDGw+sKFC5QtW5aiRYtSqVIlLC0tuXDhgro8KSmJ69ev4+LiAqCTMgyJBAFCCCF0Lj9mB1SrVg1XV1dWr17NpUuXiIiIICAggMePH9OmTRt8fX1JTk5m5cqV3Lp1i6NHj7J79246dOgAPL1u36JFCwIDAzlz5gw3b95kyZIlFC9eHC8vLwCdlGFIjJSsvg3xVjBqZnjdVSL/KMG3AEjOSNJzTURBYmFi9dpljD42Vuu8X3rP0zpvQkICW7ZsITQ0lKSkJJydnenduzcVK1YEnt5xdcOGDVy/fp2iRYvSunVrWrVqpX4+MzOTwMBAQkJCSE5OxtnZmQEDBlC6dGk1jy7KMBQSBAghhNC5McfHaZ13QYO5+VgT8TIyMFAIIYTOGefb0wOELkkQ8JaRbl/xrKxuXzkuxLN0cTkgL9f6hf5IECCEEELn5FHChkGCACGEEDpnbCSTzwyBBAFCCCF0TnoCDIMEAUIIIXTOSG5DYxAkCBBCCKFz0hNgGCQIEEIIoXMyO8AwSBAghBBC54zkPgEGQYIAIYQQOmdiLGMCDIEEAUIIIXTOWAYGGgQJAoQQQuicjAkwDBIECCGE0DkJAgyDBAFCCCF0Th4gZBgkCBBCCKFz0hNgGCQIEEIIoXMm8uwAgyBBgBBCCJ2TBwgZBgkChBBC6JxcDjAMEgQIIYTQObljoGGQIEAIIYTOyQOEDIMEAUIIIXROBgYaBgkChBBC6JyRBAEGQYIAIYQQOidjAgyDBAFCCCF0TsYEGAYJAoQQQuicTBE0DBIECCGE0Dl5doBhkCBACCGEzhkbm+i7CkILEgQIIYTQufzqCXjy5AnfffcdZ86c4dGjR5QrV47OnTvj6ekJQHR0NOvXrycsLAxzc3MaNWpEjx49MDH5JyjZt28fu3fvJjY2lsqVK9OvXz+qV6+uLtdFGYZC5nAIIYTQOSMjI61febF8+XJ+//13/Pz8mD9/PnXr1mX+/PmcP3+e9PR0Zs6cCYC/vz9+fn4cPnyYbdu2qZ8PCQlhy5YtdO/enblz52Jvb8/MmTOJj48H0EkZhkSCACGEEDpnlId/2oqLi+P06dP07dsXNzc3ypQpQ8eOHXF1deXQoUOcPHmS+/fvM3z4cCpWrIinpyc9e/Zk7969pKamArB9+3Y++OADvL29KV++PEOGDMHKyorg4GAAnZRhSCQIEEIIoXP50RNQqFAhJk6ciLOzc7Z1JSQkEBYWRuXKlbGxsVGXubi4kJKSwrVr14iLi+Pu3bu4urqqy42NjXFyciIsLAxAJ2UYEhkTIIQQQufyMiYgMTGRxMTEbOnW1tZYW1ur7y0tLXF3d9fIc+XKFc6fP0+/fv34888/KV68uMbyrPcPHz7E3NwcgBIlSmjkKVasGOHh4QDExMS8dhmGRIIAIYQQOmdspP3sgD179hAUFJQtvXPnznTt2vWFn7t16xYLFiygevXqNG3alDNnzmBra6uRx9T06WkuLS1N7c7PSstiZmZGWloaAKmpqa9dhiGRIEAIIYTO5aWbv1Xr1vj4+GRLf7YX4HkXL15kwYIFlCxZkgkTJmBqaprjiTg9PR0ACwsL9Vd8VlqWtLQ0LCwsgJxP5nktw5BIECCEEELn8jLg7/lu/9z88ssvrFy5EhcXF0aPHo2lpSUAdnZ2XL9+XSNvTEwM8LRL387OTk2rWLGimic2Nlbt8tdFGYZEBgYKIYTQufyaInjs2DECAgKoX78+48ePVwMAAGdnZ27cuKExvuDChQtYWlpStWpVbG1tsbe35+LFi+ryzMxMLl26hIuLi87KMCTSEyCEEELn8uNmQQ8fPmT16tW4urrSq1cvEhIS1GWmpqbUqVOH4sWLs3jxYnr16sX9+/fZunUrrVu3Vq/ht2nThg0bNmBvb0/16tXZuXMnycnJNGnSBEAnZRgSI0VRFH1XQrw5yRlJ+q6CKEAsTKwAOS6Epqzj4nX8eON7rfO2q9xFq3x79+5l48aNOS5zdHTE39+fqKgo1q1bx8WLF7GxscHX15euXbtibPxPx/euXbvYu3cvjx8/pmrVqvTt25eqVauqy3VRhqGQIOAtI1/24lkSBIic6CII2Hkz+2j/F2lbqfNrr0+8GrkcIIQQQufyMjBQ6I8EAUIIIXTOOI8D/oR+SBAghBBC56QnwDBIECCEEELn8jr1T+iHBAFCCCF0Li+3DRb6I0GAEEIIncuP+wQI3ZMgQAghhM7J5QDDIEGAEEIInZOBgYZBnh0g3gohh0Ko79lAIy05OZlFXy6hRZOWNKj7PgP7fUzYxUt6qqF4E86c/o1aLh4vfN25fQeAn/buo1O7LtRxr8eHLdqy9ZtAPdfc8OTXswOEbklPgPjXO/vHWSaO+5znb445f84C9uzay8jRI6hQsQKbN2xmUP+P+d+O7yldprSeaivyk7OLE18HbtJIS0lJZczIz3BycaJM2TLs+2k/Ez6bSJ++vRk74TNOnzzN3FnzsLGxpm37tnqqueExkYGBBkGCAPGvlZqaypavt7J86QosLS3JzMxUl2VmZrJn11569+1F94+6AeDuUQufBo3Zt3c//+nfR1/VFvnIxsYGt1puGmnzZs/HyAhmz5uJkZERixcspmuPrvz3s1EA1POqy53bd/j1xCkJAvJALgcYBgkCxL/WsaPHWffVekaNGUl8XDybN36tLsvMzCQtLQ0baxs1zdLSEnNzc+Lj4/VRXaEH4VfD+Xbrd0z4fDzFixfnwvkL3L0bRecuHTXyzZ4/S081NFzSzW8YZEyA+NdyrenK3gN76Nn7o2xfSKampnTu2onALYGcP3eBR/GPWPzlEpKTk2na3PAeBypezbIly6lUqSKd/v+kf+Xy3wBkZGTQv88A3nWrQ/PGLfgucJs+q2mQjPLwT+iP9ASIf63SpUu9dPngYX6c++scPbv1Ap7+cvGfPR0XV5c3UT2hZ7du3ebI4SNMnva5+ojY2NhYTExM+HTYSLp174rfUD8OHTzELP/ZFClahBYtP9BzrQ2H9AQYBgkCCoghQ4ZonXflypX5WJO3w5MnT+jzUV/SUlOZMcefUqVK8XPwz0z9fBo21tb4NvHVdxVFPvvh+x+wtbWlzYet1bT09HQyMjLo1KUjA/0GAE/HBNy+fYfVy1dLEJAHxtLRbBAkCCggevXqxcqVKylXrhyenp76rs6/3s/Bh4i4GcGW776hRk1X4OmXfVxcHHNmzpMg4C1w+OfD+DbxwdzcXE2zsrICoIG35nTS+u958eW86BOVqAAAEJFJREFUhaSlpmFmbvZG62mojI0kCDAEEgQUEA0aNMDMzIwlS5YwdOhQKlWqpO8q/avdi4rCxMQE1xqaXf8etT3Y/9MBkhKTsLK20lPtRH67e+cu165dV2cAZKlQsQIAaWlpGunpaekoioKRsXRxa0suBxgGCdUKkLp16/Lee+8RGCg3JslvlSpXIiMjg3N/ndNIP/fXeYoVL4allaWeaibehPPnzgNQs1ZNjfR3PWtTqFAhgvcHa6QfPfoLrjVcMTWV303akoGBhkGO6AKmf//+REdH67sa/3qNfBvh6OTIZ6PGMXzEUEqWLMmRkKPs2bWH8ZPGya+Yf7mrf4dTrFhRihYtqpFuY2PDgI/7s2r5aqxtrPH0fJf9+w4QeiaU5auW6am2hklO7oZBgoACxtLSUi4FvAFmZmasWb+KRQsW8+W8hSQnp1C1WhUWLJpHsw+a6bt6Ip/FxMRQuHDhHJf5DfkYGxsbvt3yLZvWb6ZS5Up8uWQBDd5vkGN+8QISSBsEI+X5e6mKf7XkjCR9V0EUIBYmT8c9yHEhnpV1XLyO0Ae/ap33Xbv6r70+8WqkJ0AIIYTOyewAwyBBgBBCCJ2TMQGGQYIAIYQQOieDaw2DBAFCCCF0TnoCDIMEAQXElClTtM47bdq0fKyJEEK8PgkCDIMEAQVE6dKl9V0FIYTQmTc1MHD79u38/vvv+Pv7q2nR0dGsX7+esLAwzM3NadSoET169MDExETNs2/fPnbv3k1sbCyVK1emX79+VK9eXadlGAIJAgqIoUOH6rsKQgihM29iTMD+/fv59ttvcXBwUNPS09OZOXMmZcuWxd/fn+joaFauXImJiQk9evQAICQkhC1btuDn50flypX58ccfmTlzJosXL6ZIkSI6KcNQyByOAiouLo6LFy9y/vx5zp8/z7lz5/j999/Ztk2eay6EKPjy87bBMTExzJkzh2+++QZ7e3uNZSdPnuT+/fsMHz6cihUr4unpSc+ePdm7dy+pqanA096DDz74AG9vb8qXL8+QIUOwsrIiODhYZ2UYCukJKICOHDnCmjVrSE9Pz7asVKlSdO3aVQ+1EkII7eXnmIBr165haWnJggULCAoKIioqSl0WFhZG5cqVsbGxUdNcXFxISUnh2rVrlClThrt37+Lq6qouNzY2xsnJibCwMJ2VYSgkCCiAduzYQYMGDejQoQMTJ05k8uTJxMTEsG7dOrp06aLv6gkhRK7ycjkgMTGRxMTEbOnW1tZYW1tnS/f09HzhI9djYmIoXry4RlrW+4cPH6qPji5RooRGnmLFihEeHq6zMgyFBAEFUHR0NGPGjKFs2bJUrlyZR48e4enpSUZGBtu3b6dhw4b6rqIQQrxUXnoC9uzZQ1BQULb0zp0757nnMzU1FVtbW420rKc/pqWlqd35zz8R0szMTH2EtC7KMBQSBBRApqammJmZAU9nDURERODu7k61atW4e/eunmsnhBC5y8vsgNatW+Pj45MtPadegNzkdCLOurRqYWGh/op//nJrWloaFhYWOivDUMjAwAKoWrVqHDhwAICKFSvy559/AhARESHPMxdCGAgjrV/W1taUKlUq2+tVggC7/2vv/mOirh8/gD/fyMHdeRwcPz4i11eFkETKgZ3DMcqKtowodAtsZT/Imi0scqt/WuQa6mqWwlLH3JCNqdhiNjEpYBJlmCtjtSG1z0fhwoS3Ht3xAbzPyR133z8cF8gRB3d47zf3fGxs8uJ9r/eL28vjea8f94qNhcVimVBmNpsB3BrSj42NnVA2xmKxuIf8/VGHXDAESFBBQQEaGxtRX1+P7OxsdHV1oaSkBOXl5VizZk2gm0dENC1BELz+8qfU1FQYjcYJawwuXrwIlUqFpKQkaLVaJCQkoLOz0/1zp9OJ33//HStXrvRbHXLBt5USlJqaioqKCjgcDmi1WpSVlaG5uRkxMTHIzc0NdPOIiKYVqE8MXLNmDaKjo1FeXo7NmzfDZDLh2LFjeOKJJ9wjqXl5eaiurkZCQgKSk5NRX18Pm82GnJwcv9UhF4LL5XIFuhF05/DceBpv7Nx49gsab6xf+KJ76N9eX5sYkTL9RVM4cOAARFGc8ImBoiiiqqoKnZ2d0Gg0ePjhh1FYWIiQkL8Hv0+dOoWGhgYMDQ0hKSkJL730EpKSkvxahxwwBEjQdGcDzOScgdvxxZ7GYwggT/wRAozD//H62mWa5T7fj2aH0wESdPvCktHRUYiiiKtXr+Lxxx8PUKuIiLwXwiVnssAQIEFvvPGGx/Ljx49jeHj4DreGiGjm7sTZAeQ7RjUZeeSRR9DW1hboZhARTWsuzw4g/+FIgIxcuXIl0E0gIvIKRwLkgSFAgvbv3z+pzGq1oqOjA2vXrg1Ai4iIZobv8OWBIUCC+vv7J6Xo0NBQPPXUU8jLywtQq4iIvMcQIA/cIigzAwMDiIqKmvXjuRWMxuMWQfLEH1sExf95P30Zr/o/n+9Hs8OFgRK0adMmDA4OTirv7++fcucAEZG0eH92AAUOpwMk4ty5c/jll1/c3x8+fNh9UtUYk8kkuxOqiCg48U+7PDAESERKSgoaGxsxNjtjMpmwYMEC988FQYBSqURxcXGgmkhENAOMAXLANQES9MEHH+Dtt9+e1TGa0+HcL43HNQHkiT/WBFy39Xp97b+UCT7fj2aHIUCiLly4AEEQcP/99wMAKisrkZmZiYyMDJ/q5Ys9jccQQJ74IwSYbH1eXxunXOzz/Wh2uDBQglpaWrB3716YTCZ3mcPhwJ49e3Du3LkAtoyIyDv8xEB54JoACaqvr8eWLVsmnEu9bds2rFixAidOnEBWVlYAW0dERPMFRwIkqL+/H/fdd9+k8lWrVkEUxQC0iIhoZgRB8PqLAochQILi4uLQ0dExqfy3336DTqcLQIuIiGg+4nSABD322GOorq6GKIpITk4GAFy+fBlfffUVCgoKAtw6IqLpca5fHhgCJGj9+vUYGRnB6dOncfLkSQCATqfDc889h8zMzAC3johoegwB8sAtghI3NDSE0NBQdHV1oampCRcuXMDRo0dnXR+3gtF43CJInvhji+DASL/X10aFxfp8P5odjgRImNVqxdmzZ9Hc3Ize3l4oFAqsW7cu0M0iIvICRwLkgCFAgi5fvozm5ma0tbVhZGQEAJCbm4uNGzdCq9UGuHVERNNjBJAHhgCJGBkZwffff4/m5mZ0dXVBqVQiMzMTWVlZ2LNnD3JychgAiEhGGAPkgCFAIrZu3QqXy4X09HRs2LABq1evhkKhCHSziIhmhfv/5YEhQCKcTifUajXCw8PhcDgwOjrKEEBEssXdAfLAECARhw4dwg8//IBvvvkGra2tCAsLQ0ZGBj8imIhkiiFADrhFUIL6+vrQ0tKC7777DgMDAwCAdevWIT8/H3q93qe6uRWMxuMWQfLEH1sEh+3/9fpajSLS5/vR7DAESJjT6UR7eztaW1vR3t6O0dFR3HvvvSgtLZ11nXyxp/EYAsgTf4SAG45Br69dGMpFz4HC6QAJCwkJgcFggMFgwODgIFpbW9Ha2hroZhEReYHTAXLAkYAgw3d8NB5HAsgTf4wEWB3DXl+rDtX4fD+aHY4EEBGR33GLoDwwBAQZfyR8mn/YL4iCE6cDiIiIglRIoBtAREREgcEQQEREFKQYAoiIiIIUQwAREVGQYgggIiIKUgwBREREQYohgIiIKEgxBBAREQUphgAiIqIgxY8NpoArLi6GyWRyfy8IAlQqFe6++24UFhbinnvu8ev9RFHEm2++iR07diAtLQ0HDhyAKIooKyub9rEulwvffvstMjIyEBnp+Qx0p9OJuro6nDlzBjdu3MCKFSvwyiuvID4+3q+/x3w23/rEeF988QXa29u9qptorjEEkCTk5uZiw4YNAG69qA4NDaG2thY7d+7Evn37EBsbO2f3LioqgtPp9Orajo4OHDx4EPv375/ymrq6OjQ1NeH1119HTEwMjh07hl27duGTTz5BWFiYv5o9782nPjGmsbERx48fR0pKiq9NJPILTgeQJCiVSkRFRSEqKgo6nQ5LlizBq6++ips3b+LHH3+c03ur1WpoNN4dZTrdURt2ux1ffvklCgoKsHr1aixduhQlJSUYGBjA+fPn/dHcoDFf+gQAmM1mfPjhhzhy5AgSEhJ8bR6R33AkgCRrwYIFAACFQgHg1hBxZmYmfv31V5jNZpSUlCA9PR0tLS04deoUrl+/jri4ODz00EN48skn3Y/v6elBdXU1Ll26BJ1Oh/z8/An3uX3o9/r166ipqUFHRwcEQUBaWhqKioogiiJ27twJANi2bRuefvppFBYWTqjLaDTCZrMhLS3NXaZWq5GYmIjOzk48+OCDc/NkBQk59gkA6Orqgkqlwscff4y6ujqIojhnzxHRTDAEkCSZzWbU1NRAqVQiIyPDXf7111/jnXfeQWRkJPR6PZqamlBbW4stW7Zg+fLluHLlCqqqqnDt2jVs3boVVqsVZWVlSE5Oxq5du2CxWFBZWTnlfa1WK95//33o9XqUlpYiNDQUhw8fxkcffYTdu3dj+/bt2LdvH3bv3o277rrLY7sBICYmZkK5TqfDX3/95adnJzjJtU8AgMFggMFg8PtzQuQrhgCShJMnT+L06dMAbi2ss9vt0Ov12L59+4S531WrVk34A3DixAls3LgR2dnZAIBFixbBbrejvLwczz77LM6fPw+bzYbi4mJoNBosWbIEL774Ivbu3euxHW1tbRgeHsZbb72FiIgIAMBrr72GM2fOYGRkBAsXLgQAaLVaKJXKSY+/efMmgL/fqY5RKBSw2+2zfXqC0nzpE0RSxhBAkpCTk4O8vDwAQEhICDQaDdRq9aTrFi9e7P734OAgzGYzPvvsM3z++efucpfLBZfLhb6+PvT09GDRokUT5nf/aVFWT08P4uPj3S/2Y/fcvHmzV7/H2MI/u92O0NC//3vZ7Xb+gZih+dIniKSMIYAkQaPReLWFLjw83P3vsdXbzz//PNLT0yddGx0djbNnz05auDX+j/Pt/uln3hh7h2qxWKBSqdzlFotlyqFi8my+9AkiKePuAJKtyMhIREZG4tq1a4iPj3d/Xb16FUePHoXD4UBiYiJEUcTg4KD7cZcuXZqyTr1eD1EUMTw87C7r7e1FUVER/vzzTwiC8I9tWrp0KVQqFS5evOgus1qt6O7uxsqVK334bckbUuwTRFLGEECyJQgC8vPz0djYiIaGBoiiiJ9//tm9yEutViMrKwtRUVGoqKiA0WhEZ2cnampqpqzzgQceQEREBD799FMYjUZ0d3ejsrIS8fHx0Ov17nf3RqMRVqt10uMVCgXWr1+P2tpa/PTTT/jjjz9QUVGB6OhorF27dm6eCHKTYp8gkjKOc5Gs5eXlITw8HA0NDThy5Ai0Wi2ys7PxzDPPALi113zHjh2oqqpCaWkpIiIiUFhYiIMHD3qsLzw8HO+99x5qampQWloKhUKB9PR0vPDCCxAEAcuWLYPBYEB5eTkeffRRvPzyy5Pq2LRpE5xOJw4dOgSbzYbU1FS8++67HFa+Q6TYJ4ikSnB580kXRERENO9wOoCIiChIMQQQEREFKYYAIiKiIMUQQEREFKQYAoiIiIIUQwAREVGQYgggIiIKUgwBREREQYohgIiIKEj9Pz6DIfAKpE7NAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9979284435237527\n",
      "Sensitivity: 0.8085106382978723\n",
      "Specificity: 0.9982415418161356\n",
      "Precision: 0.4318181818181818\n",
      "f1 score: 0.562962962962963\n",
      "AUC value is: 0.9758707735270481\n",
      "Model classification metrics have finished calculating!\n",
      "{'accuracy': 0.9979284435237527, 'sensitivity': 0.8085106382978723, 'specificity': 0.9982415418161356, 'precision': 0.4318181818181818, 'f1': 0.562962962962963, 'auc': 0.9758707735270481, 'pr_auc': 0.7139050836796361}\n",
      "Model fitting and results are complete!\n",
      "Accuracy with all fraud results is 73.57723577235772%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1440 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# finalcolstouse is list of features that are significant (p < 0.05) for logistic regression + response\n",
    "# modelpipeline.run_model(self, df, varlist, response, testratio, standardize, sampletype, modelname, text, CV)\n",
    "results = modelpipeline.run_model(df, finalcolstouse, 'Class', 0.2, True, 'naive', 'RandomForest', 'Random Forest with backward linear selection', False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount', 'Class']\n",
      "Oversampling is complete!\n",
      "[[56773    95]\n",
      " [   18    76]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9980162213405428\n",
      "Sensitivity: 0.8085106382978723\n",
      "Specificity: 0.9983294647253288\n",
      "Precision: 0.4444444444444444\n",
      "f1 score: 0.5735849056603773\n",
      "AUC value is: 0.9767051993492957\n",
      "Model classification metrics have finished calculating!\n",
      "{'accuracy': 0.9980162213405428, 'sensitivity': 0.8085106382978723, 'specificity': 0.9983294647253288, 'precision': 0.4444444444444444, 'f1': 0.5735849056603773, 'auc': 0.9767051993492957, 'pr_auc': 0.6345816120820706}\n",
      "Model fitting and results are complete!\n",
      "Accuracy with all fraud results is 75.0%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1440 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Based on backward selection and naive sampling, V18 was removed\n",
    "finalcolstouse = list(df.columns.values)\n",
    "print(finalcolstouse)\n",
    "# Compare to using all features\n",
    "# run_model(self, df, varlist, response, testratio, standardize, sampletype, modelname, text, CV)\n",
    "results = modelpipeline.run_model(df, finalcolstouse, 'Class', 0.2, True, 'naive', 'RandomForest', 'Random Forest with backward linear selection', False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# XGBoost Model with and without V18 feature removed from backward selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V18', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount', 'Class']\n",
      "Oversampling is complete!\n",
      "[0]\tvalidation_0-error:0.024964\tvalidation_1-error:0.011025\n",
      "[5]\tvalidation_0-error:0.016786\tvalidation_1-error:0.007321\n",
      "[10]\tvalidation_0-error:0.009457\tvalidation_1-error:0.005003\n",
      "[15]\tvalidation_0-error:0.006377\tvalidation_1-error:0.003652\n",
      "[20]\tvalidation_0-error:0.001139\tvalidation_1-error:0.002949\n",
      "[25]\tvalidation_0-error:0.000864\tvalidation_1-error:0.00244\n",
      "[30]\tvalidation_0-error:0.000662\tvalidation_1-error:0.001931\n",
      "[35]\tvalidation_0-error:0.000534\tvalidation_1-error:0.001668\n",
      "[40]\tvalidation_0-error:0.000446\tvalidation_1-error:0.001492\n",
      "[45]\tvalidation_0-error:0.000396\tvalidation_1-error:0.001369\n",
      "[50]\tvalidation_0-error:0.000336\tvalidation_1-error:0.001317\n",
      "[55]\tvalidation_0-error:0.000299\tvalidation_1-error:0.001194\n",
      "[60]\tvalidation_0-error:0.000273\tvalidation_1-error:0.001141\n",
      "[65]\tvalidation_0-error:0.000233\tvalidation_1-error:0.000983\n",
      "[70]\tvalidation_0-error:0.000163\tvalidation_1-error:0.000808\n",
      "[75]\tvalidation_0-error:0.000136\tvalidation_1-error:0.000737\n",
      "[80]\tvalidation_0-error:0.000112\tvalidation_1-error:0.000685\n",
      "[85]\tvalidation_0-error:0.000103\tvalidation_1-error:0.000667\n",
      "[90]\tvalidation_0-error:7.9e-05\tvalidation_1-error:0.00065\n",
      "[95]\tvalidation_0-error:7.5e-05\tvalidation_1-error:0.00065\n",
      "[99]\tvalidation_0-error:6.2e-05\tvalidation_1-error:0.000632\n",
      "[[56852    16]\n",
      " [   20    74]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9993679997191109\n",
      "Sensitivity: 0.7872340425531915\n",
      "Specificity: 0.9997186466905817\n",
      "Precision: 0.8222222222222222\n",
      "f1 score: 0.8043478260869565\n",
      "AUC value is: 0.9819135092988766\n",
      "Model classification metrics have finished calculating!\n",
      "{'accuracy': 0.9993679997191109, 'sensitivity': 0.7872340425531915, 'specificity': 0.9997186466905817, 'precision': 0.8222222222222222, 'f1': 0.8043478260869565, 'auc': 0.9819135092988766, 'pr_auc': 0.7943091970136639}\n",
      "Model fitting and results are complete!\n",
      "Accuracy with all fraud results is 79.0650406504065%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1440 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Based on backward selection and naive sampling, V18 was removed\n",
    "finalcolstouse = list(df.columns.values)\n",
    "print(finalcolstouse)\n",
    "# finalcolstouse is list of features that are significant (p < 0.05) for logistic regression + response\n",
    "# run_model(self, df, varlist, response, testratio, standardize, sampletype, modelname, text, CV)\n",
    "results = modelpipeline.run_model(df, finalcolstouse, 'Class', 0.2, True, 'naive', 'XGBoost', 'XGBoost with all features', False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Remove V18 and rebuild XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['V1', 'V2', 'V3', 'V4', 'V5', 'V6', 'V7', 'V8', 'V9', 'V10', 'V11', 'V12', 'V13', 'V14', 'V15', 'V16', 'V17', 'V19', 'V20', 'V21', 'V22', 'V23', 'V24', 'V25', 'V26', 'V27', 'V28', 'Amount', 'Class']\n",
      "Oversampling is complete!\n",
      "[0]\tvalidation_0-error:0.024964\tvalidation_1-error:0.011025\n",
      "[5]\tvalidation_0-error:0.017595\tvalidation_1-error:0.009076\n",
      "[10]\tvalidation_0-error:0.012185\tvalidation_1-error:0.005109\n",
      "[15]\tvalidation_0-error:0.002836\tvalidation_1-error:0.003178\n",
      "[20]\tvalidation_0-error:0.001277\tvalidation_1-error:0.002739\n",
      "[25]\tvalidation_0-error:0.001046\tvalidation_1-error:0.002317\n",
      "[30]\tvalidation_0-error:0.000802\tvalidation_1-error:0.001861\n",
      "[35]\tvalidation_0-error:0.000666\tvalidation_1-error:0.001668\n",
      "[40]\tvalidation_0-error:0.000594\tvalidation_1-error:0.001404\n",
      "[45]\tvalidation_0-error:0.000459\tvalidation_1-error:0.001282\n",
      "[50]\tvalidation_0-error:0.000393\tvalidation_1-error:0.001141\n",
      "[55]\tvalidation_0-error:0.000343\tvalidation_1-error:0.001088\n",
      "[60]\tvalidation_0-error:0.000299\tvalidation_1-error:0.001001\n",
      "[65]\tvalidation_0-error:0.000264\tvalidation_1-error:0.000948\n",
      "[70]\tvalidation_0-error:0.0002\tvalidation_1-error:0.00079\n",
      "[75]\tvalidation_0-error:0.000158\tvalidation_1-error:0.000755\n",
      "[80]\tvalidation_0-error:0.000141\tvalidation_1-error:0.000702\n",
      "[85]\tvalidation_0-error:0.000114\tvalidation_1-error:0.000667\n",
      "[90]\tvalidation_0-error:9.7e-05\tvalidation_1-error:0.000614\n",
      "[95]\tvalidation_0-error:9.2e-05\tvalidation_1-error:0.000597\n",
      "[99]\tvalidation_0-error:8.8e-05\tvalidation_1-error:0.000597\n",
      "[[56855    13]\n",
      " [   21    73]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.999403110845827\n",
      "Sensitivity: 0.776595744680851\n",
      "Specificity: 0.9997714004360976\n",
      "Precision: 0.8488372093023255\n",
      "f1 score: 0.8111111111111111\n",
      "AUC value is: 0.982341899643669\n",
      "Model classification metrics have finished calculating!\n",
      "{'accuracy': 0.999403110845827, 'sensitivity': 0.776595744680851, 'specificity': 0.9997714004360976, 'precision': 0.8488372093023255, 'f1': 0.8111111111111111, 'auc': 0.982341899643669, 'pr_auc': 0.7894768640900444}\n",
      "Model fitting and results are complete!\n",
      "Accuracy with all fraud results is 78.86178861788618%\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1440 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Based on backward selection and naive sampling, V18 was removed\n",
    "finalcolstouse = list(df.columns.values)\n",
    "finalcolstouse.remove('V18')\n",
    "print(finalcolstouse)\n",
    "# run_model(self, df, varlist, response, testratio, standardize, sampletype, modelname, text, CV)\n",
    "results = modelpipeline.run_model(df, finalcolstouse, 'Class', 0.2, True, 'naive', 'XGBoost', 'XGBoost with backward logistic selection', False)"
   ]
  },
  {
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}