FeatureSelection.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature selection"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Use the diabetes dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "names = ['preg', 'plas', 'pres', 'skin', 'test', 'mass', 'pedi', 'age', 'class']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_diabetes = pd.read_csv('diabetes.csv', names = names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "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>preg</th>\n",
       "      <th>plas</th>\n",
       "      <th>pres</th>\n",
       "      <th>skin</th>\n",
       "      <th>test</th>\n",
       "      <th>mass</th>\n",
       "      <th>pedi</th>\n",
       "      <th>age</th>\n",
       "      <th>class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>148</td>\n",
       "      <td>72</td>\n",
       "      <td>35</td>\n",
       "      <td>0</td>\n",
       "      <td>33.6</td>\n",
       "      <td>0.627</td>\n",
       "      <td>50</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>85</td>\n",
       "      <td>66</td>\n",
       "      <td>29</td>\n",
       "      <td>0</td>\n",
       "      <td>26.6</td>\n",
       "      <td>0.351</td>\n",
       "      <td>31</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>183</td>\n",
       "      <td>64</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>23.3</td>\n",
       "      <td>0.672</td>\n",
       "      <td>32</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>66</td>\n",
       "      <td>23</td>\n",
       "      <td>94</td>\n",
       "      <td>28.1</td>\n",
       "      <td>0.167</td>\n",
       "      <td>21</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>137</td>\n",
       "      <td>40</td>\n",
       "      <td>35</td>\n",
       "      <td>168</td>\n",
       "      <td>43.1</td>\n",
       "      <td>2.288</td>\n",
       "      <td>33</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   preg  plas  pres  skin  test  mass   pedi  age  class\n",
       "0     6   148    72    35     0  33.6  0.627   50      1\n",
       "1     1    85    66    29     0  26.6  0.351   31      0\n",
       "2     8   183    64     0     0  23.3  0.672   32      1\n",
       "3     1    89    66    23    94  28.1  0.167   21      0\n",
       "4     0   137    40    35   168  43.1  2.288   33      1"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_diabetes.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df_diabetes[['preg', 'plas', 'pres', 'skin', 'test', 'mass', 'pedi', 'age']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y = df_diabetes['class']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1) Remove features with low variance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_selection import VarianceThreshold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "var = VarianceThreshold(threshold=0.3)\n",
    "var = var.fit(X,Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2, 3, 4, 5, 7], dtype=int32)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols = var.get_support(indices=True)\n",
    "cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['preg', 'plas', 'pres', 'skin', 'test', 'mass', 'age'], dtype='object')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features = X.columns[cols]\n",
    "features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pearson Correlation Filters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x1200 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,12))\n",
    "cor = df_diabetes.corr()\n",
    "sns.heatmap(cor, annot=True, cmap=plt.cm.Reds)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "preg     0.221898\n",
       "plas     0.466581\n",
       "test     0.130548\n",
       "mass     0.292695\n",
       "pedi     0.173844\n",
       "age      0.238356\n",
       "class    1.000000\n",
       "Name: class, dtype: float64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Consider correlations only with the target variable\n",
    "cor_target = abs(cor['class'])\n",
    "\n",
    "#Select correlations with a correlation above a threshold\n",
    "features = cor_target[cor_target>0.1]\n",
    "features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['preg', 'plas', 'test', 'mass', 'pedi', 'age', 'class'], dtype='object')"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features.index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## KBest Fit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_selection import SelectKBest\n",
    "from sklearn.feature_selection import chi2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "KBest = SelectKBest(score_func = chi2, k = 5)\n",
    "KBest = KBest.fit(X,Y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Scores and p values for all features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>features</th>\n",
       "      <th>Chi2Score</th>\n",
       "      <th>pValue</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>preg</td>\n",
       "      <td>111.519691</td>\n",
       "      <td>4.552610e-26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>plas</td>\n",
       "      <td>1411.887041</td>\n",
       "      <td>5.487286e-309</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>pres</td>\n",
       "      <td>17.605373</td>\n",
       "      <td>2.718193e-05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>skin</td>\n",
       "      <td>53.108040</td>\n",
       "      <td>3.156977e-13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>test</td>\n",
       "      <td>2175.565273</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>mass</td>\n",
       "      <td>127.669343</td>\n",
       "      <td>1.325908e-29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>pedi</td>\n",
       "      <td>5.392682</td>\n",
       "      <td>2.022137e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>age</td>\n",
       "      <td>181.303689</td>\n",
       "      <td>2.516388e-41</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  features    Chi2Score         pValue\n",
       "0     preg   111.519691   4.552610e-26\n",
       "1     plas  1411.887041  5.487286e-309\n",
       "2     pres    17.605373   2.718193e-05\n",
       "3     skin    53.108040   3.156977e-13\n",
       "4     test  2175.565273   0.000000e+00\n",
       "5     mass   127.669343   1.325908e-29\n",
       "6     pedi     5.392682   2.022137e-02\n",
       "7      age   181.303689   2.516388e-41"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_scores = pd.DataFrame({'features': X.columns, 'Chi2Score': KBest.scores_, 'pValue': KBest.pvalues_ })\n",
    "df_scores"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Selected features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 4, 5, 7], dtype=int32)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols = KBest.get_support(indices=True)\n",
    "cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['preg', 'plas', 'test', 'mass', 'age'], dtype='object')"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features = X.columns[cols]\n",
    "features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SelectPercentile"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_selection import SelectPercentile\n",
    "from sklearn.feature_selection import chi2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "SPercentile = SelectPercentile(score_func = chi2, percentile=80)\n",
    "SPercentile = SPercentile.fit(X,Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": true
   },
   "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>features</th>\n",
       "      <th>Chi2Score</th>\n",
       "      <th>pValue</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>preg</td>\n",
       "      <td>111.519691</td>\n",
       "      <td>4.552610e-26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>plas</td>\n",
       "      <td>1411.887041</td>\n",
       "      <td>5.487286e-309</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>pres</td>\n",
       "      <td>17.605373</td>\n",
       "      <td>2.718193e-05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>skin</td>\n",
       "      <td>53.108040</td>\n",
       "      <td>3.156977e-13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>test</td>\n",
       "      <td>2175.565273</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>mass</td>\n",
       "      <td>127.669343</td>\n",
       "      <td>1.325908e-29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>pedi</td>\n",
       "      <td>5.392682</td>\n",
       "      <td>2.022137e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>age</td>\n",
       "      <td>181.303689</td>\n",
       "      <td>2.516388e-41</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  features    Chi2Score         pValue\n",
       "0     preg   111.519691   4.552610e-26\n",
       "1     plas  1411.887041  5.487286e-309\n",
       "2     pres    17.605373   2.718193e-05\n",
       "3     skin    53.108040   3.156977e-13\n",
       "4     test  2175.565273   0.000000e+00\n",
       "5     mass   127.669343   1.325908e-29\n",
       "6     pedi     5.392682   2.022137e-02\n",
       "7      age   181.303689   2.516388e-41"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_scores = pd.DataFrame({'features': X.columns, 'Chi2Score': SPercentile.scores_, 'pValue': SPercentile.pvalues_ })\n",
    "df_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 3, 4, 5, 7], dtype=int32)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols = SPercentile.get_support(indices=True)\n",
    "cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['preg', 'plas', 'skin', 'test', 'mass', 'age'], dtype='object')"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features = X.columns[cols]\n",
    "features"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Sequential Feature Selection (StepWise Forward Selector)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from mlxtend.feature_selection import SequentialFeatureSelector as sfs\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Estimator\n",
    "estimator = RandomForestClassifier(n_estimators=10, n_jobs=-1)\n",
    "# Step Forward Feature Selector\n",
    "StepForward = sfs(estimator,k_features=6,forward=True,floating=False,verbose=2,scoring='accuracy',cv=5)                                   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:   17.3s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:  2.2min finished\n",
      "\n",
      "[2019-12-12 14:54:52] Features: 1/6 -- score: 0.7162295221118751[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:   15.1s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   7 out of   7 | elapsed:  1.8min finished\n",
      "\n",
      "[2019-12-12 14:56:38] Features: 2/6 -- score: 0.7148374501315677[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:   15.0s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   6 out of   6 | elapsed:  1.5min finished\n",
      "\n",
      "[2019-12-12 14:58:08] Features: 3/6 -- score: 0.7435022493846024[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:   15.0s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   5 out of   5 | elapsed:  1.3min finished\n",
      "\n",
      "[2019-12-12 14:59:24] Features: 4/6 -- score: 0.7526271114506409[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:   15.2s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:  1.0min finished\n",
      "\n",
      "[2019-12-12 15:00:25] Features: 5/6 -- score: 0.7578473813767932[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:   15.5s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:   45.2s finished\n",
      "\n",
      "[2019-12-12 15:01:10] Features: 6/6 -- score: 0.7474238180120534"
     ]
    },
    {
     "data": {
      "text/plain": [
       "SequentialFeatureSelector(clone_estimator=True, cv=5,\n",
       "                          estimator=RandomForestClassifier(bootstrap=True,\n",
       "                                                           class_weight=None,\n",
       "                                                           criterion='gini',\n",
       "                                                           max_depth=None,\n",
       "                                                           max_features='auto',\n",
       "                                                           max_leaf_nodes=None,\n",
       "                                                           min_impurity_decrease=0.0,\n",
       "                                                           min_impurity_split=None,\n",
       "                                                           min_samples_leaf=1,\n",
       "                                                           min_samples_split=2,\n",
       "                                                           min_weight_fraction_leaf=0.0,\n",
       "                                                           n_estimators=10,\n",
       "                                                           n_jobs=-1,\n",
       "                                                           oob_score=False,\n",
       "                                                           random_state=None,\n",
       "                                                           verbose=0,\n",
       "                                                           warm_start=False),\n",
       "                          floating=False, forward=True, k_features=6, n_jobs=1,\n",
       "                          pre_dispatch='2*n_jobs', scoring='accuracy',\n",
       "                          verbose=2)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "StepForward.fit(X,Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{1: {'feature_idx': (1,),\n",
       "  'cv_scores': array([0.69480519, 0.7012987 , 0.68831169, 0.74509804, 0.75163399]),\n",
       "  'avg_score': 0.7162295221118751,\n",
       "  'feature_names': ('plas',)},\n",
       " 2: {'feature_idx': (1, 7),\n",
       "  'cv_scores': array([0.74675325, 0.71428571, 0.68831169, 0.69934641, 0.7254902 ]),\n",
       "  'avg_score': 0.7148374501315677,\n",
       "  'feature_names': ('plas', 'age')},\n",
       " 3: {'feature_idx': (1, 5, 7),\n",
       "  'cv_scores': array([0.72077922, 0.73376623, 0.76623377, 0.77777778, 0.71895425]),\n",
       "  'avg_score': 0.7435022493846024,\n",
       "  'feature_names': ('plas', 'mass', 'age')},\n",
       " 4: {'feature_idx': (1, 5, 6, 7),\n",
       "  'cv_scores': array([0.77922078, 0.72077922, 0.74025974, 0.78431373, 0.73856209]),\n",
       "  'avg_score': 0.7526271114506409,\n",
       "  'feature_names': ('plas', 'mass', 'pedi', 'age')},\n",
       " 5: {'feature_idx': (0, 1, 5, 6, 7),\n",
       "  'cv_scores': array([0.74675325, 0.72727273, 0.77272727, 0.80392157, 0.73856209]),\n",
       "  'avg_score': 0.7578473813767932,\n",
       "  'feature_names': ('preg', 'plas', 'mass', 'pedi', 'age')},\n",
       " 6: {'feature_idx': (0, 1, 2, 5, 6, 7),\n",
       "  'cv_scores': array([0.74025974, 0.73376623, 0.74675325, 0.78431373, 0.73202614]),\n",
       "  'avg_score': 0.7474238180120534,\n",
       "  'feature_names': ('preg', 'plas', 'pres', 'mass', 'pedi', 'age')}}"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "StepForward.subsets_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0, 1, 2, 5, 6, 7]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Selected feature columns\n",
    "cols = list(StepForward.k_feature_idx_)\n",
    "cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['preg', 'plas', 'pres', 'mass', 'pedi', 'age'], dtype='object')"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features = X.columns[cols]\n",
    "features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('preg', 'plas', 'pres', 'mass', 'pedi', 'age')"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "StepForward.k_feature_names_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7474238180120534"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "StepForward.k_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Sequential Feature Selection (StepWise Backward Eliminator)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "from mlxtend.feature_selection import SequentialFeatureSelector as sfs\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Estimator\n",
    "estimator = RandomForestClassifier(n_estimators=10, n_jobs=-1)\n",
    "# Step Forward Feature Selector\n",
    "StepBackward = sfs(estimator,k_features='best',forward=False,floating=False,verbose=2,scoring='accuracy',cv=5)                                   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.7s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    6.2s finished\n",
      "\n",
      "[2019-09-24 15:00:21] Features: 7/1 -- score: 0.7591545709192768[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.7s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   7 out of   7 | elapsed:    5.5s finished\n",
      "\n",
      "[2019-09-24 15:00:27] Features: 6/1 -- score: 0.7721840251252015[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.7s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   6 out of   6 | elapsed:    4.6s finished\n",
      "\n",
      "[2019-09-24 15:00:32] Features: 5/1 -- score: 0.7696120872591461[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.7s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   5 out of   5 | elapsed:    3.9s finished\n",
      "\n",
      "[2019-09-24 15:00:36] Features: 4/1 -- score: 0.7461845344198285[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.7s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    3.1s finished\n",
      "\n",
      "[2019-09-24 15:00:39] Features: 3/1 -- score: 0.7383329089211441[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.7s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    2.3s finished\n",
      "\n",
      "[2019-09-24 15:00:41] Features: 2/1 -- score: 0.7174688057040999[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.7s remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    1.5s finished\n",
      "\n",
      "[2019-09-24 15:00:43] Features: 1/1 -- score: 0.7175536881419233"
     ]
    },
    {
     "data": {
      "text/plain": [
       "SequentialFeatureSelector(clone_estimator=True, cv=5,\n",
       "                          estimator=RandomForestClassifier(bootstrap=True,\n",
       "                                                           class_weight=None,\n",
       "                                                           criterion='gini',\n",
       "                                                           max_depth=None,\n",
       "                                                           max_features='auto',\n",
       "                                                           max_leaf_nodes=None,\n",
       "                                                           min_impurity_decrease=0.0,\n",
       "                                                           min_impurity_split=None,\n",
       "                                                           min_samples_leaf=1,\n",
       "                                                           min_samples_split=2,\n",
       "                                                           min_weight_fraction_leaf=0.0,\n",
       "                                                           n_estimators=10,\n",
       "                                                           n_jobs=-1,\n",
       "                                                           oob_score=False,\n",
       "                                                           random_state=None,\n",
       "                                                           verbose=0,\n",
       "                                                           warm_start=False),\n",
       "                          floating=False, forward=False, k_features='best',\n",
       "                          n_jobs=1, pre_dispatch='2*n_jobs', scoring='accuracy',\n",
       "                          verbose=2)"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "StepBackward.fit(X,Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{8: {'feature_idx': (0, 1, 2, 3, 4, 5, 6, 7),\n",
       "  'cv_scores': array([0.72727273, 0.73376623, 0.72727273, 0.77777778, 0.74509804]),\n",
       "  'avg_score': 0.7422375010610305,\n",
       "  'feature_names': ('preg',\n",
       "   'plas',\n",
       "   'pres',\n",
       "   'skin',\n",
       "   'test',\n",
       "   'mass',\n",
       "   'pedi',\n",
       "   'age')},\n",
       " 7: {'feature_idx': (0, 1, 2, 4, 5, 6, 7),\n",
       "  'cv_scores': array([0.75974026, 0.72077922, 0.76623377, 0.82352941, 0.7254902 ]),\n",
       "  'avg_score': 0.7591545709192768,\n",
       "  'feature_names': ('preg', 'plas', 'pres', 'test', 'mass', 'pedi', 'age')},\n",
       " 6: {'feature_idx': (1, 2, 4, 5, 6, 7),\n",
       "  'cv_scores': array([0.74675325, 0.75324675, 0.77922078, 0.81045752, 0.77124183]),\n",
       "  'avg_score': 0.7721840251252015,\n",
       "  'feature_names': ('plas', 'pres', 'test', 'mass', 'pedi', 'age')},\n",
       " 5: {'feature_idx': (1, 2, 4, 5, 7),\n",
       "  'cv_scores': array([0.75974026, 0.73376623, 0.75324675, 0.85620915, 0.74509804]),\n",
       "  'avg_score': 0.7696120872591461,\n",
       "  'feature_names': ('plas', 'pres', 'test', 'mass', 'age')},\n",
       " 4: {'feature_idx': (1, 2, 5, 7),\n",
       "  'cv_scores': array([0.74675325, 0.69480519, 0.72727273, 0.81699346, 0.74509804]),\n",
       "  'avg_score': 0.7461845344198285,\n",
       "  'feature_names': ('plas', 'pres', 'mass', 'age')},\n",
       " 3: {'feature_idx': (1, 5, 7),\n",
       "  'cv_scores': array([0.74675325, 0.69480519, 0.73376623, 0.78431373, 0.73202614]),\n",
       "  'avg_score': 0.7383329089211441,\n",
       "  'feature_names': ('plas', 'mass', 'age')},\n",
       " 2: {'feature_idx': (1, 5),\n",
       "  'cv_scores': array([0.69480519, 0.72077922, 0.72077922, 0.74509804, 0.70588235]),\n",
       "  'avg_score': 0.7174688057040999,\n",
       "  'feature_names': ('plas', 'mass')},\n",
       " 1: {'feature_idx': (1,),\n",
       "  'cv_scores': array([0.67532468, 0.72077922, 0.67532468, 0.75163399, 0.76470588]),\n",
       "  'avg_score': 0.7175536881419233,\n",
       "  'feature_names': ('plas',)}}"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "StepBackward.subsets_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1, 2, 4, 5, 6, 7]"
      ]
     },
     "execution_count": 132,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Selected feature columns\n",
    "cols = list(StepBackward.k_feature_idx_)\n",
    "cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['plas', 'pres', 'test', 'mass', 'pedi', 'age'], dtype='object')"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features = X.columns[cols]\n",
    "features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('plas', 'pres', 'test', 'mass', 'pedi', 'age')"
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "StepBackward.k_feature_names_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7721840251252015"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "StepBackward.k_score_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Recursive Feature Elimination"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_selection import RFE\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "estimator = RandomForestClassifier(n_estimators=10, n_jobs=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "rfe = RFE(estimator=estimator, n_features_to_select=4, step=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "rfe = rfe.fit(X, Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 1, 3, 4, 5, 1, 1, 1])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfe.ranking_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 5, 6, 7], dtype=int32)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols = rfe.get_support(indices=True)\n",
    "cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['plas', 'mass', 'pedi', 'age'], dtype='object')"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features = X.columns[cols]\n",
    "features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_RFE = df_diabetes[features]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "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>plas</th>\n",
       "      <th>mass</th>\n",
       "      <th>pedi</th>\n",
       "      <th>age</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>148</td>\n",
       "      <td>33.6</td>\n",
       "      <td>0.627</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>85</td>\n",
       "      <td>26.6</td>\n",
       "      <td>0.351</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>183</td>\n",
       "      <td>23.3</td>\n",
       "      <td>0.672</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>89</td>\n",
       "      <td>28.1</td>\n",
       "      <td>0.167</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>137</td>\n",
       "      <td>43.1</td>\n",
       "      <td>2.288</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   plas  mass   pedi  age\n",
       "0   148  33.6  0.627   50\n",
       "1    85  26.6  0.351   31\n",
       "2   183  23.3  0.672   32\n",
       "3    89  28.1  0.167   21\n",
       "4   137  43.1  2.288   33"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_RFE.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Recursive Feature Elimination with Cross Validation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### With Random Forests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_selection import RFECV\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "estimator = RandomForestClassifier(n_estimators=10, n_jobs=-1, random_state = 13)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "rfecv = RFECV(estimator, cv=5, n_jobs=-1, scoring='accuracy', min_features_to_select = 4)\n",
    "rfecv = rfecv.fit(X, Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 1, 1, 3, 4, 1, 1, 1])"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfecv.ranking_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfecv.n_features_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.75259316, 0.76956965, 0.75266106, 0.74482642, 0.75132841])"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfecv.grid_scores_\n",
    "# These are for the 5 cross validation grids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 2, 5, 6, 7], dtype=int32)"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## Best features\n",
    "cols = rfecv.get_support(indices=True)\n",
    "cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['plas', 'pres', 'mass', 'pedi', 'age'], dtype='object')"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features = X.columns[cols]\n",
    "features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_RFECV = df_diabetes[features]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "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>plas</th>\n",
       "      <th>pres</th>\n",
       "      <th>mass</th>\n",
       "      <th>pedi</th>\n",
       "      <th>age</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>148</td>\n",
       "      <td>72</td>\n",
       "      <td>33.6</td>\n",
       "      <td>0.627</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>85</td>\n",
       "      <td>66</td>\n",
       "      <td>26.6</td>\n",
       "      <td>0.351</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>183</td>\n",
       "      <td>64</td>\n",
       "      <td>23.3</td>\n",
       "      <td>0.672</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>89</td>\n",
       "      <td>66</td>\n",
       "      <td>28.1</td>\n",
       "      <td>0.167</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>137</td>\n",
       "      <td>40</td>\n",
       "      <td>43.1</td>\n",
       "      <td>2.288</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   plas  pres  mass   pedi  age\n",
       "0   148    72  33.6  0.627   50\n",
       "1    85    66  26.6  0.351   31\n",
       "2   183    64  23.3  0.672   32\n",
       "3    89    66  28.1  0.167   21\n",
       "4   137    40  43.1  2.288   33"
      ]
     },
     "execution_count": 213,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_RFECV.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### With XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [],
   "source": [
    "from xgboost import XGBClassifier\n",
    "from sklearn.feature_selection import RFECV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [],
   "source": [
    "estimator = XGBClassifier()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "rfecv = RFECV(estimator, cv=5, n_jobs=-1, scoring='accuracy', min_features_to_select = 4)\n",
    "rfecv = rfecv.fit(X, Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 1, 1, 1, 1, 1, 1, 1])"
      ]
     },
     "execution_count": 156,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfecv.ranking_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8"
      ]
     },
     "execution_count": 157,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rfecv.n_features_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Best features\n",
    "cols = rfecv.get_support(indices=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [],
   "source": [
    "col_names = X.columns[cols]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_RFECV = df_diabetes[col_names]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "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>preg</th>\n",
       "      <th>plas</th>\n",
       "      <th>pres</th>\n",
       "      <th>skin</th>\n",
       "      <th>test</th>\n",
       "      <th>mass</th>\n",
       "      <th>pedi</th>\n",
       "      <th>age</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6</td>\n",
       "      <td>148</td>\n",
       "      <td>72</td>\n",
       "      <td>35</td>\n",
       "      <td>0</td>\n",
       "      <td>33.6</td>\n",
       "      <td>0.627</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>85</td>\n",
       "      <td>66</td>\n",
       "      <td>29</td>\n",
       "      <td>0</td>\n",
       "      <td>26.6</td>\n",
       "      <td>0.351</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8</td>\n",
       "      <td>183</td>\n",
       "      <td>64</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>23.3</td>\n",
       "      <td>0.672</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>89</td>\n",
       "      <td>66</td>\n",
       "      <td>23</td>\n",
       "      <td>94</td>\n",
       "      <td>28.1</td>\n",
       "      <td>0.167</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>137</td>\n",
       "      <td>40</td>\n",
       "      <td>35</td>\n",
       "      <td>168</td>\n",
       "      <td>43.1</td>\n",
       "      <td>2.288</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   preg  plas  pres  skin  test  mass   pedi  age\n",
       "0     6   148    72    35     0  33.6  0.627   50\n",
       "1     1    85    66    29     0  26.6  0.351   31\n",
       "2     8   183    64     0     0  23.3  0.672   32\n",
       "3     1    89    66    23    94  28.1  0.167   21\n",
       "4     0   137    40    35   168  43.1  2.288   33"
      ]
     },
     "execution_count": 161,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_RFECV.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Probably XGBoost is not the best estimator for RFE or XGBoost here doesn't require feature elimination"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prinicipal Component Analysis (PCA)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = MinMaxScaler(feature_range=[0, 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_rescaled = scaler.fit_transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.35294118, 0.74371859, 0.59016393, ..., 0.50074516, 0.23441503,\n",
       "        0.48333333],\n",
       "       [0.05882353, 0.42713568, 0.54098361, ..., 0.39642325, 0.11656704,\n",
       "        0.16666667],\n",
       "       [0.47058824, 0.91959799, 0.52459016, ..., 0.34724292, 0.25362938,\n",
       "        0.18333333],\n",
       "       ...,\n",
       "       [0.29411765, 0.6080402 , 0.59016393, ..., 0.390462  , 0.07130658,\n",
       "        0.15      ],\n",
       "       [0.05882353, 0.63316583, 0.49180328, ..., 0.4485842 , 0.11571307,\n",
       "        0.43333333],\n",
       "       [0.05882353, 0.46733668, 0.57377049, ..., 0.45305514, 0.10119556,\n",
       "        0.03333333]])"
      ]
     },
     "execution_count": 224,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_rescaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_rescaled = pd.DataFrame(data = X_rescaled, columns = ['preg', 'plas', 'pres', 'skin', 'test', 'mass', 'pedi', 'age'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "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>preg</th>\n",
       "      <th>plas</th>\n",
       "      <th>pres</th>\n",
       "      <th>skin</th>\n",
       "      <th>test</th>\n",
       "      <th>mass</th>\n",
       "      <th>pedi</th>\n",
       "      <th>age</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.352941</td>\n",
       "      <td>0.743719</td>\n",
       "      <td>0.590164</td>\n",
       "      <td>0.353535</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.500745</td>\n",
       "      <td>0.234415</td>\n",
       "      <td>0.483333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.058824</td>\n",
       "      <td>0.427136</td>\n",
       "      <td>0.540984</td>\n",
       "      <td>0.292929</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.396423</td>\n",
       "      <td>0.116567</td>\n",
       "      <td>0.166667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.470588</td>\n",
       "      <td>0.919598</td>\n",
       "      <td>0.524590</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.347243</td>\n",
       "      <td>0.253629</td>\n",
       "      <td>0.183333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.058824</td>\n",
       "      <td>0.447236</td>\n",
       "      <td>0.540984</td>\n",
       "      <td>0.232323</td>\n",
       "      <td>0.111111</td>\n",
       "      <td>0.418778</td>\n",
       "      <td>0.038002</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.688442</td>\n",
       "      <td>0.327869</td>\n",
       "      <td>0.353535</td>\n",
       "      <td>0.198582</td>\n",
       "      <td>0.642325</td>\n",
       "      <td>0.943638</td>\n",
       "      <td>0.200000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       preg      plas      pres      skin      test      mass      pedi  \\\n",
       "0  0.352941  0.743719  0.590164  0.353535  0.000000  0.500745  0.234415   \n",
       "1  0.058824  0.427136  0.540984  0.292929  0.000000  0.396423  0.116567   \n",
       "2  0.470588  0.919598  0.524590  0.000000  0.000000  0.347243  0.253629   \n",
       "3  0.058824  0.447236  0.540984  0.232323  0.111111  0.418778  0.038002   \n",
       "4  0.000000  0.688442  0.327869  0.353535  0.198582  0.642325  0.943638   \n",
       "\n",
       "        age  \n",
       "0  0.483333  \n",
       "1  0.166667  \n",
       "2  0.183333  \n",
       "3  0.000000  \n",
       "4  0.200000  "
      ]
     },
     "execution_count": 226,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_rescaled.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "metadata": {},
   "outputs": [],
   "source": [
    "pca = PCA()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "metadata": {},
   "outputs": [],
   "source": [
    "pca = pca.fit(X_rescaled, Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 250,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting the Cumulative Summation of the Explained Variance\n",
    "plt.figure(figsize=(12,8))\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_))\n",
    "plt.xlabel('Number of Components')\n",
    "plt.ylabel('Variance (%)') #for each component\n",
    "plt.title('Diabetes Data Explained Variance')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Automatically find optimum number of features through numpy interpolation method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import math"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "metadata": {},
   "outputs": [],
   "source": [
    "xp = list(range(len(np.cumsum(pca.explained_variance_ratio_))))\n",
    "yp = np.cumsum(pca.explained_variance_ratio_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 264,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_components = np.interp(0.95, yp, xp) + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_components = math.ceil(n_components)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 268,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7"
      ]
     },
     "execution_count": 268,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "n_components"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Automated PCA decomposition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.decomposition import PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "metadata": {},
   "outputs": [],
   "source": [
    "pca = PCA(n_components=7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "metadata": {},
   "outputs": [],
   "source": [
    "pca = pca.fit(X, Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-2.02176587e-03,  9.78115765e-02,  1.60930503e-02,\n",
       "         6.07566861e-02,  9.93110844e-01,  1.40108085e-02,\n",
       "         5.37167919e-04, -3.56474430e-03],\n",
       "       [-2.26488861e-02, -9.72210040e-01, -1.41909330e-01,\n",
       "         5.78614699e-02,  9.46266913e-02, -4.69729766e-02,\n",
       "        -8.16804621e-04, -1.40168181e-01],\n",
       "       [-2.24649003e-02,  1.43428710e-01, -9.22467192e-01,\n",
       "        -3.07013055e-01,  2.09773019e-02, -1.32444542e-01,\n",
       "        -6.39983017e-04, -1.25454310e-01],\n",
       "       [-4.90459604e-02,  1.19830016e-01, -2.62742788e-01,\n",
       "         8.84369380e-01, -6.55503615e-02,  1.92801728e-01,\n",
       "         2.69908637e-03, -3.01024330e-01],\n",
       "       [ 1.51612874e-01, -8.79407680e-02, -2.32165009e-01,\n",
       "         2.59973487e-01, -1.72312241e-04,  2.14744823e-02,\n",
       "         1.64080684e-03,  9.20504903e-01],\n",
       "       [-5.04730888e-03,  5.07391813e-02,  7.56365525e-02,\n",
       "         2.21363068e-01, -6.13326472e-03, -9.70776708e-01,\n",
       "        -2.02903702e-03, -1.51133239e-02],\n",
       "       [ 9.86672995e-01,  8.83426114e-04, -1.22975947e-03,\n",
       "        -3.76444746e-04,  1.42307394e-03, -2.73046214e-03,\n",
       "        -6.34402965e-03, -1.62555343e-01]])"
      ]
     },
     "execution_count": 237,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca.components_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_matrix = pca.fit_transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-75.714655</td>\n",
       "      <td>-35.950783</td>\n",
       "      <td>-7.260789</td>\n",
       "      <td>15.669269</td>\n",
       "      <td>16.506541</td>\n",
       "      <td>3.460442</td>\n",
       "      <td>-0.702047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-82.358268</td>\n",
       "      <td>28.908213</td>\n",
       "      <td>-5.496671</td>\n",
       "      <td>9.004554</td>\n",
       "      <td>3.481527</td>\n",
       "      <td>5.590262</td>\n",
       "      <td>-2.572015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-74.630643</td>\n",
       "      <td>-67.906496</td>\n",
       "      <td>19.461808</td>\n",
       "      <td>-5.653056</td>\n",
       "      <td>-10.300113</td>\n",
       "      <td>7.144367</td>\n",
       "      <td>4.279067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11.077423</td>\n",
       "      <td>34.898486</td>\n",
       "      <td>-0.053018</td>\n",
       "      <td>1.314873</td>\n",
       "      <td>-7.619414</td>\n",
       "      <td>2.583855</td>\n",
       "      <td>-0.809829</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>89.743788</td>\n",
       "      <td>-2.746937</td>\n",
       "      <td>25.212859</td>\n",
       "      <td>18.994237</td>\n",
       "      <td>8.522694</td>\n",
       "      <td>-9.486986</td>\n",
       "      <td>-3.626410</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           0          1          2          3          4         5         6\n",
       "0 -75.714655 -35.950783  -7.260789  15.669269  16.506541  3.460442 -0.702047\n",
       "1 -82.358268  28.908213  -5.496671   9.004554   3.481527  5.590262 -2.572015\n",
       "2 -74.630643 -67.906496  19.461808  -5.653056 -10.300113  7.144367  4.279067\n",
       "3  11.077423  34.898486  -0.053018   1.314873  -7.619414  2.583855 -0.809829\n",
       "4  89.743788  -2.746937  25.212859  18.994237   8.522694 -9.486986 -3.626410"
      ]
     },
     "execution_count": 240,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(feature_matrix).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##  Feature Importance of Tree-based classifiers: ExtraTreesClassifier (Note Bagged Decision Trees or other trees could also be used)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import ExtraTreesClassifier\n",
    "from sklearn.feature_selection import SelectFromModel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "estimator = ExtraTreesClassifier(n_estimators=10)\n",
    "estimator = estimator.fit(X, Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.11750309, 0.21451933, 0.09232234, 0.06912186, 0.07827504,\n",
       "       0.14348476, 0.13021632, 0.15455726])"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "estimator.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = SelectFromModel(estimator, prefit=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_new = model.transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[148.   ,  33.6  ,   0.627,  50.   ],\n",
       "       [ 85.   ,  26.6  ,   0.351,  31.   ],\n",
       "       [183.   ,  23.3  ,   0.672,  32.   ],\n",
       "       ...,\n",
       "       [121.   ,  26.2  ,   0.245,  30.   ],\n",
       "       [126.   ,  30.1  ,   0.349,  47.   ],\n",
       "       [ 93.   ,  30.4  ,   0.315,  23.   ]])"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_new"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(768, 4)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_new.shape "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 5, 6, 7], dtype=int32)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols = model.get_support(indices = True)\n",
    "cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['plas', 'mass', 'pedi', 'age'], dtype='object')"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features = X.columns[cols]\n",
    "features"
   ]
  },
  {
   "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
}