FS-Wrappers.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature Selection using Wrappers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "`scikit learn` does not provide a comprehenisive implementation of Wrapper feature selection so we use `MLxtend`.  \n",
    "http://rasbt.github.io/mlxtend/\n",
    "So you will probably need to install some libraries:  \n",
    "`pip install mlxtend`  \n",
    "`pip install joblib`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from mlxtend.feature_selection import SequentialFeatureSelector as SFS\n",
    "from matplotlib.ticker import MaxNLocator"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Forward Sequential Search on segmentation data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2310, 20)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Class</th>\n",
       "      <th>REGION-CENTROID-COL</th>\n",
       "      <th>REGION-CENTROID-ROW</th>\n",
       "      <th>REGION-PIXEL-COUNT</th>\n",
       "      <th>SHORT-LINE-DENSITY-5</th>\n",
       "      <th>SHORT-LINE-DENSITY-2</th>\n",
       "      <th>VEDGE-MEAN</th>\n",
       "      <th>VEDGE-SD</th>\n",
       "      <th>HEDGE-MEAN</th>\n",
       "      <th>HEDGE-SD</th>\n",
       "      <th>INTENSITY-MEAN</th>\n",
       "      <th>RAWRED-MEAN</th>\n",
       "      <th>RAWBLUE-MEAN</th>\n",
       "      <th>RAWGREEN-MEAN</th>\n",
       "      <th>EXRED-MEAN</th>\n",
       "      <th>EXBLUE-MEAN</th>\n",
       "      <th>EXGREEN-MEAN</th>\n",
       "      <th>VALUE-MEAN</th>\n",
       "      <th>SATURATION-MEAN</th>\n",
       "      <th>HUE-MEAN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BRICKFACE</td>\n",
       "      <td>140.0</td>\n",
       "      <td>125.0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.277778</td>\n",
       "      <td>0.062963</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.311111</td>\n",
       "      <td>6.185185</td>\n",
       "      <td>7.333334</td>\n",
       "      <td>7.666666</td>\n",
       "      <td>3.555556</td>\n",
       "      <td>3.444444</td>\n",
       "      <td>4.444445</td>\n",
       "      <td>-7.888889</td>\n",
       "      <td>7.777778</td>\n",
       "      <td>0.545635</td>\n",
       "      <td>-1.121818</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BRICKFACE</td>\n",
       "      <td>188.0</td>\n",
       "      <td>133.0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.266667</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.077778</td>\n",
       "      <td>6.666666</td>\n",
       "      <td>8.333334</td>\n",
       "      <td>7.777778</td>\n",
       "      <td>3.888889</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>3.333333</td>\n",
       "      <td>-8.333333</td>\n",
       "      <td>8.444445</td>\n",
       "      <td>0.538580</td>\n",
       "      <td>-0.924817</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BRICKFACE</td>\n",
       "      <td>105.0</td>\n",
       "      <td>139.0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.277778</td>\n",
       "      <td>0.107407</td>\n",
       "      <td>0.833333</td>\n",
       "      <td>0.522222</td>\n",
       "      <td>6.111111</td>\n",
       "      <td>7.555555</td>\n",
       "      <td>7.222222</td>\n",
       "      <td>3.555556</td>\n",
       "      <td>4.333334</td>\n",
       "      <td>3.333333</td>\n",
       "      <td>-7.666666</td>\n",
       "      <td>7.555555</td>\n",
       "      <td>0.532628</td>\n",
       "      <td>-0.965946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BRICKFACE</td>\n",
       "      <td>34.0</td>\n",
       "      <td>137.0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.166667</td>\n",
       "      <td>1.111111</td>\n",
       "      <td>0.474074</td>\n",
       "      <td>5.851852</td>\n",
       "      <td>7.777778</td>\n",
       "      <td>6.444445</td>\n",
       "      <td>3.333333</td>\n",
       "      <td>5.777778</td>\n",
       "      <td>1.777778</td>\n",
       "      <td>-7.555555</td>\n",
       "      <td>7.777778</td>\n",
       "      <td>0.573633</td>\n",
       "      <td>-0.744272</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>BRICKFACE</td>\n",
       "      <td>39.0</td>\n",
       "      <td>111.0</td>\n",
       "      <td>9</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.722222</td>\n",
       "      <td>0.374074</td>\n",
       "      <td>0.888889</td>\n",
       "      <td>0.429629</td>\n",
       "      <td>6.037037</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>7.666666</td>\n",
       "      <td>3.444444</td>\n",
       "      <td>2.888889</td>\n",
       "      <td>4.888889</td>\n",
       "      <td>-7.777778</td>\n",
       "      <td>7.888889</td>\n",
       "      <td>0.562919</td>\n",
       "      <td>-1.175773</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Class  REGION-CENTROID-COL  REGION-CENTROID-ROW  REGION-PIXEL-COUNT  \\\n",
       "0  BRICKFACE                140.0                125.0                   9   \n",
       "1  BRICKFACE                188.0                133.0                   9   \n",
       "2  BRICKFACE                105.0                139.0                   9   \n",
       "3  BRICKFACE                 34.0                137.0                   9   \n",
       "4  BRICKFACE                 39.0                111.0                   9   \n",
       "\n",
       "   SHORT-LINE-DENSITY-5  SHORT-LINE-DENSITY-2  VEDGE-MEAN  VEDGE-SD  \\\n",
       "0                   0.0                   0.0    0.277778  0.062963   \n",
       "1                   0.0                   0.0    0.333333  0.266667   \n",
       "2                   0.0                   0.0    0.277778  0.107407   \n",
       "3                   0.0                   0.0    0.500000  0.166667   \n",
       "4                   0.0                   0.0    0.722222  0.374074   \n",
       "\n",
       "   HEDGE-MEAN  HEDGE-SD  INTENSITY-MEAN  RAWRED-MEAN  RAWBLUE-MEAN  \\\n",
       "0    0.666667  0.311111        6.185185     7.333334      7.666666   \n",
       "1    0.500000  0.077778        6.666666     8.333334      7.777778   \n",
       "2    0.833333  0.522222        6.111111     7.555555      7.222222   \n",
       "3    1.111111  0.474074        5.851852     7.777778      6.444445   \n",
       "4    0.888889  0.429629        6.037037     7.000000      7.666666   \n",
       "\n",
       "   RAWGREEN-MEAN  EXRED-MEAN  EXBLUE-MEAN  EXGREEN-MEAN  VALUE-MEAN  \\\n",
       "0       3.555556    3.444444     4.444445     -7.888889    7.777778   \n",
       "1       3.888889    5.000000     3.333333     -8.333333    8.444445   \n",
       "2       3.555556    4.333334     3.333333     -7.666666    7.555555   \n",
       "3       3.333333    5.777778     1.777778     -7.555555    7.777778   \n",
       "4       3.444444    2.888889     4.888889     -7.777778    7.888889   \n",
       "\n",
       "   SATURATION-MEAN  HUE-MEAN  \n",
       "0         0.545635 -1.121818  \n",
       "1         0.538580 -0.924817  \n",
       "2         0.532628 -0.965946  \n",
       "3         0.573633 -0.744272  \n",
       "4         0.562919 -1.175773  "
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "seg_data = pd.read_csv('segmentation-all.csv')\n",
    "print(seg_data.shape)\n",
    "seg_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Class\n",
       "BRICKFACE    330\n",
       "SKY          330\n",
       "FOLIAGE      330\n",
       "CEMENT       330\n",
       "WINDOW       330\n",
       "PATH         330\n",
       "GRASS        330\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "seg_data['Class'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Prep \n",
    "- Extract the data from the dataframe into numpy arrays\n",
    "- Split into train and test sets \n",
    "- Apply a [0,1] Scaler. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((1155, 19), (1155, 19))"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = seg_data['Class']\n",
    "X_raw = seg_data.drop('Class', axis=1)\n",
    "X_tr_raw, X_ts_raw, y_train, y_test = train_test_split(X_raw, y, \n",
    "                                                       random_state=2, test_size=1/2)\n",
    "scaler = MinMaxScaler()\n",
    "X_train = scaler.fit_transform(X_tr_raw)\n",
    "X_test = scaler.transform(X_ts_raw)\n",
    "max_k = X_train.shape[1]\n",
    "X_train.shape, X_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Baseline performance evaluation\n",
    "Using all features and *k*-NN:  \n",
    "- test performance on training data using cross validation,\n",
    "- test performance on test data using hold-out. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_Val on training all features: 0.943\n",
      "Hold Out testing all features: 0.935\n"
     ]
    }
   ],
   "source": [
    "kNN = KNeighborsClassifier(n_neighbors=4)\n",
    "kNN = kNN.fit(X_train,y_train)\n",
    "y_pred = kNN.predict(X_test)\n",
    "acc = accuracy_score(y_pred,y_test)\n",
    "cv_acc = cross_val_score(kNN, X_train, y_train, cv=8)\n",
    "\n",
    "print(\"X_Val on training all features: {0:.3f}\".format(cv_acc.mean())) \n",
    "print(\"Hold Out testing all features: {0:.3f}\".format(acc)) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1155, 19)"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.18577075e-01, 2.13389121e-01, 0.00000000e+00, ...,\n",
       "        1.80412362e-01, 4.19469100e-01, 2.36777797e-01],\n",
       "       [1.97628458e-01, 2.21757322e-01, 0.00000000e+00, ...,\n",
       "        3.32106028e-01, 3.10101450e-01, 1.72210580e-01],\n",
       "       [5.88932806e-01, 5.89958159e-01, 0.00000000e+00, ...,\n",
       "        1.54639185e-02, 5.37037000e-01, 1.76103104e-01],\n",
       "       ...,\n",
       "       [4.26877470e-01, 5.64853556e-01, 0.00000000e+00, ...,\n",
       "        7.36377012e-04, 1.11111110e-01, 1.61318832e-01],\n",
       "       [4.54545455e-01, 7.57322176e-01, 0.00000000e+00, ...,\n",
       "        4.77908712e-01, 3.24017170e-01, 1.73148315e-01],\n",
       "       [4.07114625e-01, 5.27196653e-01, 0.00000000e+00, ...,\n",
       "        4.71281285e-02, 7.31922400e-01, 1.67423957e-01]])"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sequential Forward Selection\n",
    "Run SFS with k_features set to (1,max_k) - this will remember the best result."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "verb = 0\n",
    "sfs_forward = SFS(kNN, \n",
    "                  k_features= (1, max_k), \n",
    "                  forward=True, \n",
    "                  floating=False, \n",
    "                  verbose=verb,\n",
    "                  scoring='accuracy',\n",
    "                  cv=10, n_jobs = -1) # No. of threads depends on the machine.\n",
    "\n",
    "sfs_forward = sfs_forward.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The indexes and names of the features from the best perfroming subset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 2, 6, 8, 10, 12, 18)"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sfs_forward.k_feature_idx_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['REGION-CENTROID-ROW',\n",
       " 'REGION-PIXEL-COUNT',\n",
       " 'VEDGE-SD',\n",
       " 'HEDGE-SD',\n",
       " 'RAWRED-MEAN',\n",
       " 'RAWGREEN-MEAN',\n",
       " 'HUE-MEAN']"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sfs_forward_feature_names = [X_raw.columns[idx] for idx in sfs_forward.k_feature_idx_]\n",
    "sfs_forward_feature_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['REGION-CENTROID-ROW', 'REGION-PIXEL-COUNT', 'VEDGE-SD', 'HEDGE-SD', 'RAWRED-MEAN', 'RAWGREEN-MEAN', 'HUE-MEAN']\n"
     ]
    }
   ],
   "source": [
    "from mlxtend.plotting import plot_sequential_feature_selection as plot_sfs\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "fig1 = plot_sfs(sfs_forward.get_metric_dict(), \n",
    "                ylabel='Training Accuracy',\n",
    "                kind='std_dev')\n",
    "\n",
    "plt.ylim([0.5, 1])\n",
    "plt.title('Sequential Forward Selection')\n",
    "plt.grid()\n",
    "plt.show()\n",
    "print(sfs_forward_feature_names)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Transform the dataset using the selected subset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train shape:  (1155, 7)\n",
      "X_Val on SFS all features: 0.961\n",
      "Hold Out testing:  7 features selected using SFS: 0.948\n"
     ]
    }
   ],
   "source": [
    "X_train_sfs = sfs_forward.transform(X_train)\n",
    "X_test_sfs = sfs_forward.transform(X_test)\n",
    "\n",
    "kNN_sfs = kNN.fit(X_train_sfs,y_train)\n",
    "y_pred = kNN_sfs.predict(X_test_sfs)\n",
    "acc_SFS = accuracy_score(y_pred,y_test)\n",
    "cv_acc_SFS = cross_val_score(kNN, X_train_sfs, y_train, cv=8)\n",
    "\n",
    "print(\"X_train shape: \", X_train_sfs.shape)\n",
    "print(\"X_Val on SFS all features: {0:.3f}\".format(cv_acc_SFS.mean())) \n",
    "print(\"Hold Out testing: {0:2d} features selected using SFS: {1:.3f}\".format(len(sfs_forward.k_feature_idx_), acc_SFS)) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Backward Elimination\n",
    "If we set the SFS `forward` parameter to False it performs Backward Elimination."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 18/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 17/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 16/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 15/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 14/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 13/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 12/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 11/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 10/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 9/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 8/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 7/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 6/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 5/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 4/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 3/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 2/1[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "Features: 1/1"
     ]
    }
   ],
   "source": [
    "verb = 1\n",
    "sfs_backward = SFS(kNN, \n",
    "                  k_features=(1, max_k), \n",
    "                  forward=False, \n",
    "                  floating=False, \n",
    "                  verbose=verb,\n",
    "                  scoring='accuracy',\n",
    "                  cv=10, n_jobs = -1)\n",
    "\n",
    "sfs_backward = sfs_backward.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(['REGION-CENTROID-ROW',\n",
       "  'REGION-PIXEL-COUNT',\n",
       "  'HEDGE-MEAN',\n",
       "  'INTENSITY-MEAN',\n",
       "  'RAWRED-MEAN',\n",
       "  'RAWBLUE-MEAN',\n",
       "  'RAWGREEN-MEAN',\n",
       "  'EXRED-MEAN',\n",
       "  'EXGREEN-MEAN',\n",
       "  'SATURATION-MEAN'],\n",
       " 10)"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sfs_backward_feature_names = [X_raw.columns[idx] for idx in sfs_backward.k_feature_idx_]\n",
    "sfs_backward_feature_names, len(sfs_backward.k_feature_idx_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('1', '2', '7', '9', '10', '11', '12', '13', '15', '17')\n"
     ]
    }
   ],
   "source": [
    "fig1 = plot_sfs(sfs_backward.get_metric_dict(), \n",
    "                ylabel='Accuracy',\n",
    "                kind='std_dev')\n",
    "\n",
    "plt.ylim([0.5, 1])\n",
    "plt.title('Backward Elimination (w. StdDev)')\n",
    "plt.grid()\n",
    "plt.show()\n",
    "print(sfs_backward.k_feature_names_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_train shape:  (1155, 10)\n",
      "X_Val on BE all features: 0.960\n",
      "Hold Out testing: 10 features selected using BE: 0.945\n"
     ]
    }
   ],
   "source": [
    "X_train_be = sfs_backward.transform(X_train)\n",
    "X_test_be = sfs_backward.transform(X_test)\n",
    "\n",
    "kNN_be = kNN.fit(X_train_be,y_train)\n",
    "y_pred = kNN_be.predict(X_test_be)\n",
    "acc_BE = accuracy_score(y_pred,y_test)\n",
    "cv_acc_BE = cross_val_score(kNN, X_train_be, y_train, cv=8)\n",
    "\n",
    "print(\"X_train shape: \", X_train_be.shape)\n",
    "print(\"X_Val on BE all features: {0:.3f}\".format(cv_acc_BE.mean())) \n",
    "print(\"Hold Out testing: {0:2d} features selected using BE: {1:.3f}\".format(len(sfs_backward.k_feature_idx_), acc_BE)) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot the overall results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "width = 0.2\n",
    "\n",
    "options = ['All', 'SFS', 'BE']\n",
    "n_feat = [X_train.shape[1], X_train_sfs.shape[1], X_train_be.shape[1]]\n",
    "accs = [acc,acc_SFS,acc_BE]\n",
    "xv = [cv_acc.mean(), cv_acc_SFS.mean(), cv_acc_BE.mean()]\n",
    "\n",
    "y_pos = np.arange(len(options))\n",
    "\n",
    "p1 = ax.bar(y_pos-width/2, xv, width, align='center', label = 'Train (X-val)',\n",
    "            color=['blue','blue','blue'],alpha=0.5)\n",
    "p2 = ax.bar(y_pos+width/2, accs , width, align='center', label = 'Test (Hold-out)',\n",
    "            color=['g','g','g'],alpha=0.5)\n",
    "\n",
    "ax.set_ylim([0.7, 1])\n",
    "ax2 = ax.twinx()\n",
    "\n",
    "p3 = ax2.plot([0,1,2],n_feat, color = 'red', label = 'Feature Count',\n",
    "              marker = 'x', ms = 10, linewidth=0)\n",
    "ax2.set_ylim([0, 20])\n",
    "\n",
    "ax.grid(axis = 'y')\n",
    "\n",
    "h1, l1 = ax.get_legend_handles_labels()\n",
    "h2, l2 = ax2.get_legend_handles_labels()\n",
    "ax2.legend(h1+h2, l1+l2, loc='lower right')\n",
    "\n",
    "ax2.yaxis.set_major_locator(MaxNLocator(integer=True))\n",
    "\n",
    "plt.xticks(y_pos, options)\n",
    "ax.set_ylabel('Accuracy')\n",
    "ax2.set_ylabel('Feature Count')\n",
    "\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}