Feature importance using model parameters (LR).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "from pylab import rcParams\n",
    "rcParams['figure.figsize']=10,6\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "columns=['id',\n",
    "'RI',\n",
    "'Na',\n",
    "'Mg', \n",
    "'Al',\n",
    "'Si',\n",
    "'K',\n",
    "'Ca',\n",
    "'Ba',\n",
    "'Fe', # a derivate var which can be ignored \n",
    "'class'] "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Attribute Information:\n",
    "1. Id number: 1 to 214\n",
    "2. RI: refractive index\n",
    "3. Na: Sodium (unit measurement: weight percent in corresponding oxide, as are attributes 4-10)\n",
    "4. Mg: Magnesium\n",
    "5. Al: Aluminum\n",
    "6. Si: Silicon\n",
    "7. K: Potassium\n",
    "8. Ca: Calcium\n",
    "9. Ba: Barium\n",
    "10. Fe: Iron\n",
    "11. Type of glass: (class attribute)\n",
    "  1. building_windows_float_processed\n",
    "  2. building_windows_non_float_processed\n",
    "  3. vehicle_windows_float_processed\n",
    "  4. vehicle_windows_non_float_processed (none in this database)\n",
    "  5. containers\n",
    "  6. tableware\n",
    "  7. headlamps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "df=pd.read_csv('./glass.csv', names=columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "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>id</th>\n",
       "      <th>RI</th>\n",
       "      <th>Na</th>\n",
       "      <th>Mg</th>\n",
       "      <th>Al</th>\n",
       "      <th>Si</th>\n",
       "      <th>K</th>\n",
       "      <th>Ca</th>\n",
       "      <th>Ba</th>\n",
       "      <th>Fe</th>\n",
       "      <th>class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1.52101</td>\n",
       "      <td>13.64</td>\n",
       "      <td>4.49</td>\n",
       "      <td>1.10</td>\n",
       "      <td>71.78</td>\n",
       "      <td>0.06</td>\n",
       "      <td>8.75</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>1.51761</td>\n",
       "      <td>13.89</td>\n",
       "      <td>3.60</td>\n",
       "      <td>1.36</td>\n",
       "      <td>72.73</td>\n",
       "      <td>0.48</td>\n",
       "      <td>7.83</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>1.51618</td>\n",
       "      <td>13.53</td>\n",
       "      <td>3.55</td>\n",
       "      <td>1.54</td>\n",
       "      <td>72.99</td>\n",
       "      <td>0.39</td>\n",
       "      <td>7.78</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>1.51766</td>\n",
       "      <td>13.21</td>\n",
       "      <td>3.69</td>\n",
       "      <td>1.29</td>\n",
       "      <td>72.61</td>\n",
       "      <td>0.57</td>\n",
       "      <td>8.22</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>1.51742</td>\n",
       "      <td>13.27</td>\n",
       "      <td>3.62</td>\n",
       "      <td>1.24</td>\n",
       "      <td>73.08</td>\n",
       "      <td>0.55</td>\n",
       "      <td>8.07</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id       RI     Na    Mg    Al     Si     K    Ca   Ba   Fe  class\n",
       "0   1  1.52101  13.64  4.49  1.10  71.78  0.06  8.75  0.0  0.0      1\n",
       "1   2  1.51761  13.89  3.60  1.36  72.73  0.48  7.83  0.0  0.0      1\n",
       "2   3  1.51618  13.53  3.55  1.54  72.99  0.39  7.78  0.0  0.0      1\n",
       "3   4  1.51766  13.21  3.69  1.29  72.61  0.57  8.22  0.0  0.0      1\n",
       "4   5  1.51742  13.27  3.62  1.24  73.08  0.55  8.07  0.0  0.0      1"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1cda4ade898>"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(x='class', data=df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_small=df.loc[df['class']<3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "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>id</th>\n",
       "      <th>RI</th>\n",
       "      <th>Na</th>\n",
       "      <th>Mg</th>\n",
       "      <th>Al</th>\n",
       "      <th>Si</th>\n",
       "      <th>K</th>\n",
       "      <th>Ca</th>\n",
       "      <th>Ba</th>\n",
       "      <th>Fe</th>\n",
       "      <th>class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1.52101</td>\n",
       "      <td>13.64</td>\n",
       "      <td>4.49</td>\n",
       "      <td>1.10</td>\n",
       "      <td>71.78</td>\n",
       "      <td>0.06</td>\n",
       "      <td>8.75</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>1.51761</td>\n",
       "      <td>13.89</td>\n",
       "      <td>3.60</td>\n",
       "      <td>1.36</td>\n",
       "      <td>72.73</td>\n",
       "      <td>0.48</td>\n",
       "      <td>7.83</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>1.51618</td>\n",
       "      <td>13.53</td>\n",
       "      <td>3.55</td>\n",
       "      <td>1.54</td>\n",
       "      <td>72.99</td>\n",
       "      <td>0.39</td>\n",
       "      <td>7.78</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>1.51766</td>\n",
       "      <td>13.21</td>\n",
       "      <td>3.69</td>\n",
       "      <td>1.29</td>\n",
       "      <td>72.61</td>\n",
       "      <td>0.57</td>\n",
       "      <td>8.22</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>1.51742</td>\n",
       "      <td>13.27</td>\n",
       "      <td>3.62</td>\n",
       "      <td>1.24</td>\n",
       "      <td>73.08</td>\n",
       "      <td>0.55</td>\n",
       "      <td>8.07</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id       RI     Na    Mg    Al     Si     K    Ca   Ba   Fe  class\n",
       "0   1  1.52101  13.64  4.49  1.10  71.78  0.06  8.75  0.0  0.0      1\n",
       "1   2  1.51761  13.89  3.60  1.36  72.73  0.48  7.83  0.0  0.0      1\n",
       "2   3  1.51618  13.53  3.55  1.54  72.99  0.39  7.78  0.0  0.0      1\n",
       "3   4  1.51766  13.21  3.69  1.29  72.61  0.57  8.22  0.0  0.0      1\n",
       "4   5  1.51742  13.27  3.62  1.24  73.08  0.55  8.07  0.0  0.0      1"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_small.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1cda45169e8>"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(df_small['RI'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1cda4fafcc0>"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(df_small['Fe'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1cda5568e48>"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(df_small['K'], bins=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1cda54a62e8>"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(df_small['Mg'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 271,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import SGDClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import log_loss\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.calibration import CalibratedClassifierCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train, x_test, y_train, y_test = train_test_split(df_small.drop(['id', 'class'], axis=1), df_small['class'], stratify=df_small['class'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler=MinMaxScaler()\n",
    "x_train=scaler.fit_transform(x_train)\n",
    "x_test=scaler.transform(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 284,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "alpha=[10**x for x in np.arange(-5, 40, dtype=float)]\n",
    "log_alpha=np.arange(-5,40)\n",
    "\n",
    "cv_err=[]\n",
    "train_err=[]\n",
    "for a in alpha:\n",
    "    clf=LogisticRegression(C=a, penalty='l2', class_weight='balanced')\n",
    "    clf.fit(x_train, y_train)\n",
    "    sig_clf=CalibratedClassifierCV(clf)\n",
    "    sig_clf.fit(x_train, y_train)\n",
    "    \n",
    "    predict_y=sig_clf.predict_proba(x_test)\n",
    "    cv_err.append(log_loss(y_test, predict_y))\n",
    "    \n",
    "    predict_y=sig_clf.predict_proba(x_train)\n",
    "    train_err.append(log_loss(y_train, predict_y))\n",
    "\n",
    "plt.plot(log_alpha, cv_err, label='cv error', c='b')\n",
    "plt.plot(log_alpha, train_err, label='train error', c='r')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 285,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1e+26"
      ]
     },
     "execution_count": 285,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "alpha[np.argmin(cv_err)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 286,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5480284100512239"
      ]
     },
     "execution_count": 286,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.min(cv_err)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 287,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7027027027027027"
      ]
     },
     "execution_count": 293,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_test, clf.predict(x_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1e+26, class_weight='balanced', dual=False,\n",
       "          fit_intercept=True, intercept_scaling=1, max_iter=100,\n",
       "          multi_class='warn', n_jobs=None, penalty='l2', random_state=None,\n",
       "          solver='warn', tol=0.0001, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 292,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf=LogisticRegression(C=alpha[np.argmin(cv_err)], penalty='l2', class_weight='balanced')\n",
    "clf.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 294,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  2.2788908 ],\n",
       "       [-21.10823284],\n",
       "       [-30.39377841],\n",
       "       [ -3.20900923],\n",
       "       [-28.11562127],\n",
       "       [ -4.10588636],\n",
       "       [-53.55308351],\n",
       "       [-24.22635827],\n",
       "       [  1.02587204]])"
      ]
     },
     "execution_count": 294,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.coef_.T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 295,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([['RI'],\n",
       "       ['Na'],\n",
       "       ['Mg'],\n",
       "       ['Al'],\n",
       "       ['Si'],\n",
       "       ['K'],\n",
       "       ['Ca'],\n",
       "       ['Ba'],\n",
       "       ['Fe']], dtype='<U2')"
      ]
     },
     "execution_count": 295,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array([columns[1:-1]]).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 296,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_importance=pd.DataFrame(np.hstack((np.array([columns[1:-1]]).T, clf.coef_.T)), columns=['feature', 'importance'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 306,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_importance['importance']=pd.to_numeric(feature_importance['importance'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 307,
   "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>feature</th>\n",
       "      <th>importance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>RI</td>\n",
       "      <td>2.278891</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Fe</td>\n",
       "      <td>1.025872</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Al</td>\n",
       "      <td>-3.209009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>K</td>\n",
       "      <td>-4.105886</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Na</td>\n",
       "      <td>-21.108233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Ba</td>\n",
       "      <td>-24.226358</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Si</td>\n",
       "      <td>-28.115621</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Mg</td>\n",
       "      <td>-30.393778</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Ca</td>\n",
       "      <td>-53.553084</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  feature  importance\n",
       "0      RI    2.278891\n",
       "8      Fe    1.025872\n",
       "3      Al   -3.209009\n",
       "5       K   -4.105886\n",
       "1      Na  -21.108233\n",
       "7      Ba  -24.226358\n",
       "4      Si  -28.115621\n",
       "2      Mg  -30.393778\n",
       "6      Ca  -53.553084"
      ]
     },
     "execution_count": 307,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_importance.sort_values(by='importance', ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Si content, Ca content, Mg content are most valuable in determining whether glass is float processed or non float processed"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}