6_interpretation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 7. Interpreting a Tree Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>nclaims</th>\n",
       "      <th>amount</th>\n",
       "      <th>avg</th>\n",
       "      <th>exp</th>\n",
       "      <th>coverage</th>\n",
       "      <th>fuel</th>\n",
       "      <th>use</th>\n",
       "      <th>fleet</th>\n",
       "      <th>sex</th>\n",
       "      <th>ageph</th>\n",
       "      <th>bm</th>\n",
       "      <th>agec</th>\n",
       "      <th>power</th>\n",
       "      <th>pc</th>\n",
       "      <th>town</th>\n",
       "      <th>long</th>\n",
       "      <th>lat</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1618.001036</td>\n",
       "      <td>1618.001036</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>TPL</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>private</td>\n",
       "      <td>N</td>\n",
       "      <td>male</td>\n",
       "      <td>50</td>\n",
       "      <td>5</td>\n",
       "      <td>12</td>\n",
       "      <td>77</td>\n",
       "      <td>1000</td>\n",
       "      <td>BRUSSEL</td>\n",
       "      <td>4.355223</td>\n",
       "      <td>50.845386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>PO</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>private</td>\n",
       "      <td>N</td>\n",
       "      <td>female</td>\n",
       "      <td>64</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>66</td>\n",
       "      <td>1000</td>\n",
       "      <td>BRUSSEL</td>\n",
       "      <td>4.355223</td>\n",
       "      <td>50.845386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>TPL</td>\n",
       "      <td>diesel</td>\n",
       "      <td>private</td>\n",
       "      <td>N</td>\n",
       "      <td>male</td>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>70</td>\n",
       "      <td>1000</td>\n",
       "      <td>BRUSSEL</td>\n",
       "      <td>4.355223</td>\n",
       "      <td>50.845386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>TPL</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>private</td>\n",
       "      <td>N</td>\n",
       "      <td>male</td>\n",
       "      <td>77</td>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>57</td>\n",
       "      <td>1000</td>\n",
       "      <td>BRUSSEL</td>\n",
       "      <td>4.355223</td>\n",
       "      <td>50.845386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>155.974606</td>\n",
       "      <td>155.974606</td>\n",
       "      <td>0.046575</td>\n",
       "      <td>TPL</td>\n",
       "      <td>gasoline</td>\n",
       "      <td>private</td>\n",
       "      <td>N</td>\n",
       "      <td>female</td>\n",
       "      <td>28</td>\n",
       "      <td>9</td>\n",
       "      <td>7</td>\n",
       "      <td>70</td>\n",
       "      <td>1000</td>\n",
       "      <td>BRUSSEL</td>\n",
       "      <td>4.355223</td>\n",
       "      <td>50.845386</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   nclaims       amount          avg       exp coverage      fuel      use  \\\n",
       "0        1  1618.001036  1618.001036  1.000000      TPL  gasoline  private   \n",
       "1        0     0.000000          NaN  1.000000       PO  gasoline  private   \n",
       "2        0     0.000000          NaN  1.000000      TPL    diesel  private   \n",
       "3        0     0.000000          NaN  1.000000      TPL  gasoline  private   \n",
       "4        1   155.974606   155.974606  0.046575      TPL  gasoline  private   \n",
       "\n",
       "  fleet     sex  ageph  bm  agec  power    pc     town      long        lat  \n",
       "0     N    male     50   5    12     77  1000  BRUSSEL  4.355223  50.845386  \n",
       "1     N  female     64   5     3     66  1000  BRUSSEL  4.355223  50.845386  \n",
       "2     N    male     60   0    10     70  1000  BRUSSEL  4.355223  50.845386  \n",
       "3     N    male     77   0    15     57  1000  BRUSSEL  4.355223  50.845386  \n",
       "4     N  female     28   9     7     70  1000  BRUSSEL  4.355223  50.845386  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(r'.\\data\\PC_data.txt', sep = '\\s+')\n",
    "df.columns = map(str.lower, df.columns)\n",
    "df = df.drop('id', axis = 1)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# converting type(str) columns to categorical\n",
    "df['coverage'] = df['coverage'].astype('category').cat.codes\n",
    "df['fleet'] = df['fleet'].astype('category').cat.codes\n",
    "df['sex'] = df['sex'].astype('category').cat.codes\n",
    "df['town'] = df['town'].astype('category').cat.codes\n",
    "df['fuel'] = df['fuel'].astype('category').cat.codes\n",
    "df['use'] = df['use'].astype('category').cat.codes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "exp = df['exp']\n",
    "y = np.array(df['nclaims']/df['exp'])\n",
    "x = np.array(df[['ageph','agec','bm','power','coverage','fuel','sex','fleet','use']])\n",
    "col_names = ['ageph','agec','bm','power','coverage','fuel','sex','fleet','use']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import metrics\n",
    "from sklearn import tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "rgf_dt = tree.DecisionTreeRegressor(criterion=\"mse\", max_depth = 3,\n",
    "                                      max_features = 'auto',\n",
    "                                      min_samples_split = 10,\n",
    "                                      min_samples_leaf = 10,\n",
    "                                      ccp_alpha = 0, splitter=\"best\").fit(x,y,exp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "|--- bm <= 6.50\n",
      "|   |--- bm <= 1.50\n",
      "|   |   |--- ageph <= 55.50\n",
      "|   |   |   |--- value: [0.12]\n",
      "|   |   |--- ageph >  55.50\n",
      "|   |   |   |--- value: [0.09]\n",
      "|   |--- bm >  1.50\n",
      "|   |   |--- ageph <= 57.50\n",
      "|   |   |   |--- value: [0.15]\n",
      "|   |   |--- ageph >  57.50\n",
      "|   |   |   |--- value: [0.12]\n",
      "|--- bm >  6.50\n",
      "|   |--- bm <= 10.50\n",
      "|   |   |--- ageph <= 26.50\n",
      "|   |   |   |--- value: [0.24]\n",
      "|   |   |--- ageph >  26.50\n",
      "|   |   |   |--- value: [0.19]\n",
      "|   |--- bm >  10.50\n",
      "|   |   |--- agec <= 6.50\n",
      "|   |   |   |--- value: [0.23]\n",
      "|   |   |--- agec >  6.50\n",
      "|   |   |   |--- value: [0.29]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "text_representation = tree.export_text(rgf_dt, feature_names = col_names)\n",
    "print(text_representation)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Method ```feature_importances_()``` provides the importance of a feature as the (normalized) total reduction of the criterion brought by that feature. It is also known as the Gini importance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature: ageph, Score: 0.08305980025319501\n",
      "Feature: agec, Score: 0.017491585276182237\n",
      "Feature: bm, Score: 0.8994486144706227\n",
      "Feature: power, Score: 0.0\n",
      "Feature: coverage, Score: 0.0\n",
      "Feature: fuel, Score: 0.0\n",
      "Feature: sex, Score: 0.0\n",
      "Feature: fleet, Score: 0.0\n",
      "Feature: use, Score: 0.0\n"
     ]
    }
   ],
   "source": [
    "# get importance\n",
    "importance = rgf_dt.feature_importances_\n",
    "# summarize feature importance\n",
    "for i,(v,s) in enumerate(zip(col_names, importance)):\n",
    "    print(f'Feature: {v}, Score: {s}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 7.1 One-way Partial Dependence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.inspection import plot_partial_dependence\n",
    "df_x = df[['ageph','agec','bm','power','coverage','fuel','sex','fleet','use']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "ax.set_title(\"Decision Tree\")\n",
    "tree_disp = plot_partial_dependence(rgf_dt, df_x, ['ageph','bm'], kind='average', ax=ax, line_kw={'color': 'red'})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 7.2 Two-way Partial Dependence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "fig, ax = plt.subplots(figsize=(12, 6))\n",
    "ax.set_title(\"Decision Tree\")\n",
    "tree_disp = plot_partial_dependence(rgf_dt, df_x, [('ageph','bm')], kind='average', ax=ax, line_kw={'color': 'red'})"
   ]
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}