Preprocessing-.4.scikit-learn/preProcessing-categoric_polynomial_interaction_term_4_scikit_OLS.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "Earlier I said that if you need to get to categoric data, interactions, p-values and those stuff, don’t sweat on preprocessing/munging for scikit-learn LinearRegression to consume,  and rather just use statsmodel. It might be construed that I take on the preprocessing lightly. \n",
    "\n",
    "\n",
    "Far from it, coming from the vision and signal processing background I know the importance of the task too well. It’s just that there are other precious things in life like dating the other sex, watching your children grow up, spending time with your loved ones before you lose them.  You know what I am saying.\n",
    "\n",
    "\n",
    "Some thoughts on preprocessing.\n",
    "\n",
    "\n",
    "From the beginning when a group of photons strike rods and cones in the retina triggering a change in membrane potential, the signal goes through many stages of processing before we identify an object whether it’s a cat or a dog. And most of these stages are preprocessing.  Preprocessing amplifies and estimates to tease out true signal from noise, separates objects from background, construct features, and succesively build up more abstract things until it throws these well processed and almost done things to a classifier/regressor.  It’s the preprocessing stages that are doing the heavy lifting, and those possibly machine-learning-inspired classifiers are just final make-up.\n",
    "\n",
    "\n",
    "Abstraction.  It’s everywhere.  In fact, there’s no clear distinction between preprocessing and final decision making classifier/regressor.  A classifier is just a final leg of the continuum of abstraction.  And when there is an abstraction, estimation raises its head to fill the holes to give credibility for a guessing job cause you can’t do things entirely in bottom-up fashion.  \n",
    "\n",
    "\n",
    "Scikit-learn uses the term ‘estimator’ for good reasons.  Classifiers/Regressors and any serious preprocessors/filters are estimators.  When I read “an estimator is a Python object that implements the methods fit(X, y) and predict(T)” in the sense of software interface/protocol, you could have seen silly grin on my face. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A little preprocessing. \n",
    "ISLR 3장에 있는 카테고리형 변수의 변환, 변수를 non-linear 변환하여 linear regression을 적용하는 법, 변수간 interaction을 주는 연습을 한다\n",
    "\n",
    "\n",
    "We will trace exactly as written in ISLR Ch.3 to tranform a categorical variable, polynomialize a variable, give  interaction between variables, such that the transformed form is amenable for a scikit-learn estimator to consume. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# 패키지 imports\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import statsmodels.formula.api as smf    \n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns \n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Carseats = pd.read_csv(\"../Data/carseats.csv\", index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Sales</th>\n",
       "      <th>CompPrice</th>\n",
       "      <th>Income</th>\n",
       "      <th>Advertising</th>\n",
       "      <th>Population</th>\n",
       "      <th>Price</th>\n",
       "      <th>ShelveLoc</th>\n",
       "      <th>Age</th>\n",
       "      <th>Education</th>\n",
       "      <th>Urban</th>\n",
       "      <th>US</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9.50</td>\n",
       "      <td>138</td>\n",
       "      <td>73</td>\n",
       "      <td>11</td>\n",
       "      <td>276</td>\n",
       "      <td>120</td>\n",
       "      <td>Bad</td>\n",
       "      <td>42</td>\n",
       "      <td>17</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>11.22</td>\n",
       "      <td>111</td>\n",
       "      <td>48</td>\n",
       "      <td>16</td>\n",
       "      <td>260</td>\n",
       "      <td>83</td>\n",
       "      <td>Good</td>\n",
       "      <td>65</td>\n",
       "      <td>10</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10.06</td>\n",
       "      <td>113</td>\n",
       "      <td>35</td>\n",
       "      <td>10</td>\n",
       "      <td>269</td>\n",
       "      <td>80</td>\n",
       "      <td>Medium</td>\n",
       "      <td>59</td>\n",
       "      <td>12</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7.40</td>\n",
       "      <td>117</td>\n",
       "      <td>100</td>\n",
       "      <td>4</td>\n",
       "      <td>466</td>\n",
       "      <td>97</td>\n",
       "      <td>Medium</td>\n",
       "      <td>55</td>\n",
       "      <td>14</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Yes</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>4.15</td>\n",
       "      <td>141</td>\n",
       "      <td>64</td>\n",
       "      <td>3</td>\n",
       "      <td>340</td>\n",
       "      <td>128</td>\n",
       "      <td>Bad</td>\n",
       "      <td>38</td>\n",
       "      <td>13</td>\n",
       "      <td>Yes</td>\n",
       "      <td>No</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Sales  CompPrice  Income  Advertising  Population  Price ShelveLoc  Age  \\\n",
       "1   9.50        138      73           11         276    120       Bad   42   \n",
       "2  11.22        111      48           16         260     83      Good   65   \n",
       "3  10.06        113      35           10         269     80    Medium   59   \n",
       "4   7.40        117     100            4         466     97    Medium   55   \n",
       "5   4.15        141      64            3         340    128       Bad   38   \n",
       "\n",
       "   Education Urban   US  \n",
       "1         17   Yes  Yes  \n",
       "2         10   Yes  Yes  \n",
       "3         12   Yes  Yes  \n",
       "4         14   Yes  Yes  \n",
       "5         13   Yes   No  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Carseats.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## statsmodels를 쓴 예,"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"simpletable\">\n",
       "<tr>\n",
       "                 <td></td>                    <th>coef</th>     <th>std err</th>      <th>t</th>      <th>P>|t|</th> <th>[95.0% Conf. Int.]</th> \n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Intercept</th>                       <td>   14.2992</td> <td>    5.187</td> <td>    2.757</td> <td> 0.006</td> <td>    4.102    24.497</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ShelveLoc[T.Good]</th>               <td>    4.4826</td> <td>    0.452</td> <td>    9.914</td> <td> 0.000</td> <td>    3.594     5.372</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ShelveLoc[T.Medium]</th>             <td>    1.8654</td> <td>    0.361</td> <td>    5.168</td> <td> 0.000</td> <td>    1.156     2.575</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Advertising</th>                     <td>    0.0869</td> <td>    0.034</td> <td>    2.588</td> <td> 0.010</td> <td>    0.021     0.153</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ShelveLoc[T.Good]:Advertising</th>   <td>    0.0271</td> <td>    0.048</td> <td>    0.569</td> <td> 0.570</td> <td>   -0.067     0.121</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>ShelveLoc[T.Medium]:Advertising</th> <td>    0.0069</td> <td>    0.040</td> <td>    0.173</td> <td> 0.863</td> <td>   -0.071     0.085</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Age</th>                             <td>   -0.0540</td> <td>    0.017</td> <td>   -3.180</td> <td> 0.002</td> <td>   -0.087    -0.021</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Education</th>                       <td>   -0.0332</td> <td>    0.041</td> <td>   -0.814</td> <td> 0.416</td> <td>   -0.113     0.047</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Income</th>                          <td>    0.0052</td> <td>    0.013</td> <td>    0.401</td> <td> 0.689</td> <td>   -0.020     0.031</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>CompPrice</th>                       <td>   -0.1253</td> <td>    0.080</td> <td>   -1.574</td> <td> 0.116</td> <td>   -0.282     0.031</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>np.power(CompPrice, 2)</th>          <td>    0.0005</td> <td>    0.000</td> <td>    1.674</td> <td> 0.095</td> <td> -9.3e-05     0.001</td>\n",
       "</tr>\n",
       "<tr>\n",
       "  <th>Age:Income</th>                      <td>    0.0002</td> <td>    0.000</td> <td>    0.914</td> <td> 0.361</td> <td>   -0.000     0.001</td>\n",
       "</tr>\n",
       "</table>"
      ],
      "text/plain": [
       "<class 'statsmodels.iolib.table.SimpleTable'>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lm_Carseats = smf.ols(formula = \"Sales ~ Advertising+Age+Education+Income +CompPrice + \\\n",
    "                                 np.power(CompPrice, 2) + ShelveLoc + Age:Income + ShelveLoc:Advertising\", \n",
    "                      data=Carseats).fit() \n",
    "\n",
    "lm_Carseats.summary().tables[1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 앞에 categoric variable, polynomial transformation, interaction 들을 scikit-learn OLS estimator가 하도록 하려면, \n",
    "\n",
    "\n",
    "### Feature Preprocessing  \n",
    "- scikit-learn에 있는 [Preprocessing](http://scikit-learn.org/stable/modules/preprocessing.html) 말고 ISLR의 내용 따라 함    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Carseats2 = Carseats.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Sales', 'CompPrice', 'Income', 'Advertising', 'Population', 'Price',\n",
       "       'ShelveLoc', 'Age', 'Education', 'Urban', 'US'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Carseats2.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Sales</th>\n",
       "      <th>CompPrice</th>\n",
       "      <th>Income</th>\n",
       "      <th>Advertising</th>\n",
       "      <th>Age</th>\n",
       "      <th>Education</th>\n",
       "      <th>ShelveLoc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9.50</td>\n",
       "      <td>138</td>\n",
       "      <td>73</td>\n",
       "      <td>11</td>\n",
       "      <td>42</td>\n",
       "      <td>17</td>\n",
       "      <td>Bad</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>11.22</td>\n",
       "      <td>111</td>\n",
       "      <td>48</td>\n",
       "      <td>16</td>\n",
       "      <td>65</td>\n",
       "      <td>10</td>\n",
       "      <td>Good</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Sales  CompPrice  Income  Advertising  Age  Education ShelveLoc\n",
       "1   9.50        138      73           11   42         17       Bad\n",
       "2  11.22        111      48           16   65         10      Good"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Carseats2 = Carseats2.loc[:, ['Sales', 'CompPrice', 'Income', 'Advertising', 'Age', 'Education', 'ShelveLoc']]\n",
    "Carseats2.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Sales</th>\n",
       "      <th>CompPrice</th>\n",
       "      <th>Income</th>\n",
       "      <th>Advertising</th>\n",
       "      <th>Age</th>\n",
       "      <th>Education</th>\n",
       "      <th>ShelveLoc_Good</th>\n",
       "      <th>ShelveLoc_Medium</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9.50</td>\n",
       "      <td>138</td>\n",
       "      <td>73</td>\n",
       "      <td>11</td>\n",
       "      <td>42</td>\n",
       "      <td>17</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>11.22</td>\n",
       "      <td>111</td>\n",
       "      <td>48</td>\n",
       "      <td>16</td>\n",
       "      <td>65</td>\n",
       "      <td>10</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10.06</td>\n",
       "      <td>113</td>\n",
       "      <td>35</td>\n",
       "      <td>10</td>\n",
       "      <td>59</td>\n",
       "      <td>12</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7.40</td>\n",
       "      <td>117</td>\n",
       "      <td>100</td>\n",
       "      <td>4</td>\n",
       "      <td>55</td>\n",
       "      <td>14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>4.15</td>\n",
       "      <td>141</td>\n",
       "      <td>64</td>\n",
       "      <td>3</td>\n",
       "      <td>38</td>\n",
       "      <td>13</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Sales  CompPrice  Income  Advertising  Age  Education  ShelveLoc_Good  \\\n",
       "1   9.50        138      73           11   42         17             0.0   \n",
       "2  11.22        111      48           16   65         10             1.0   \n",
       "3  10.06        113      35           10   59         12             0.0   \n",
       "4   7.40        117     100            4   55         14             0.0   \n",
       "5   4.15        141      64            3   38         13             0.0   \n",
       "\n",
       "   ShelveLoc_Medium  \n",
       "1               0.0  \n",
       "2               0.0  \n",
       "3               1.0  \n",
       "4               1.0  \n",
       "5               0.0  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Carseats2 = pd.get_dummies(Carseats2, drop_first=True)   # map() 메소드를 써서 해도 됨   \n",
    "Carseats2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Sales', 'CompPrice', 'Income', 'Advertising', 'Age', 'Education',\n",
       "       'ShelveLoc_Good', 'ShelveLoc_Medium'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Carseats2.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Sales</th>\n",
       "      <th>CompPrice</th>\n",
       "      <th>Income</th>\n",
       "      <th>Advertising</th>\n",
       "      <th>Age</th>\n",
       "      <th>Education</th>\n",
       "      <th>ShelveLoc_Good</th>\n",
       "      <th>ShelveLoc_Medium</th>\n",
       "      <th>CompPrice_sq</th>\n",
       "      <th>Age*Income</th>\n",
       "      <th>ShelveLoc_Good*Advertising</th>\n",
       "      <th>ShelveLoc_Medium*Advertising</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9.50</td>\n",
       "      <td>138</td>\n",
       "      <td>73</td>\n",
       "      <td>11</td>\n",
       "      <td>42</td>\n",
       "      <td>17</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19044</td>\n",
       "      <td>3066</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>11.22</td>\n",
       "      <td>111</td>\n",
       "      <td>48</td>\n",
       "      <td>16</td>\n",
       "      <td>65</td>\n",
       "      <td>10</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>12321</td>\n",
       "      <td>3120</td>\n",
       "      <td>16.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Sales  CompPrice  Income  Advertising  Age  Education  ShelveLoc_Good  \\\n",
       "1   9.50        138      73           11   42         17             0.0   \n",
       "2  11.22        111      48           16   65         10             1.0   \n",
       "\n",
       "   ShelveLoc_Medium  CompPrice_sq  Age*Income  ShelveLoc_Good*Advertising  \\\n",
       "1               0.0         19044        3066                         0.0   \n",
       "2               0.0         12321        3120                        16.0   \n",
       "\n",
       "   ShelveLoc_Medium*Advertising  \n",
       "1                           0.0  \n",
       "2                           0.0  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Carseats2['CompPrice_sq'] = np.power(Carseats2.CompPrice, 2)\n",
    "Carseats2['Age*Income'] = Carseats2.Age * Carseats2.Income\n",
    "Carseats2['ShelveLoc_Good*Advertising'] = Carseats2.ShelveLoc_Good * Carseats2.Advertising\n",
    "Carseats2['ShelveLoc_Medium*Advertising'] = Carseats2.ShelveLoc_Medium * Carseats2.Advertising\n",
    "Carseats2.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = Carseats2.loc[:, 'CompPrice':] \n",
    "y = Carseats2.Sales"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "intercept 14.2992015741\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[('CompPrice', -0.12533253564712765),\n",
       " ('Income', 0.0051896584127176602),\n",
       " ('Advertising', 0.086904680461086778),\n",
       " ('Age', -0.053955023401916456),\n",
       " ('Education', -0.033226187003566311),\n",
       " ('ShelveLoc_Good', 4.4826130416522503),\n",
       " ('ShelveLoc_Medium', 1.8653783686161103),\n",
       " ('CompPrice_sq', 0.00053288587332844025),\n",
       " ('Age*Income', 0.00021253940130102429),\n",
       " ('ShelveLoc_Good*Advertising', 0.027095576071671807),\n",
       " ('ShelveLoc_Medium*Advertising', 0.0068851766080135785)]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "lm_sci = LinearRegression()\n",
    "lm_sci.fit(X, y)\n",
    "\n",
    "# intercept and coefficients\n",
    "print(\"intercept\", lm_sci.intercept_)\n",
    "list(zip(X.columns, lm_sci.coef_))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- statsmodels의 결과와 같음.  Both the statsmodels & scikit-learn results are the same.\n",
    "\n",
    "\n",
    "### Note that,\n",
    "- Transform categorical variables first, then do interactions on them as needed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>400.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>7.496325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1.918581</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>3.407655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>6.071667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>7.252025</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>8.688959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>12.879883</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                0\n",
       "count  400.000000\n",
       "mean     7.496325\n",
       "std      1.918581\n",
       "min      3.407655\n",
       "25%      6.071667\n",
       "50%      7.252025\n",
       "75%      8.688959\n",
       "max     12.879883"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Some statistics of y_predict & residuals on training set\n",
    "y_predict = lm_sci.predict(X)\n",
    "lm_sci_resid = y - y_predict \n",
    "pd.DataFrame(y_predict).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Residues</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>4.000000e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.110223e-16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>2.072360e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-6.307441e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>-1.480765e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>-7.805546e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.472994e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>5.989242e+00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Residues\n",
       "count  4.000000e+02\n",
       "mean   1.110223e-16\n",
       "std    2.072360e+00\n",
       "min   -6.307441e+00\n",
       "25%   -1.480765e+00\n",
       "50%   -7.805546e-02\n",
       "75%    1.472994e+00\n",
       "max    5.989242e+00"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lm_sci_resid.name = 'Residues'\n",
    "pd.DataFrame(lm_sci_resid).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0xa027eb0>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmwAAAIICAYAAADT8LJpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/x+DTHptx4GiqAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x9ffabf0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Residual Plot on training set\n",
    "fig, ax1 = plt.subplots(nrows=1, ncols=1, figsize=(10, 8))\n",
    "# sns.set_context(font_scale=1.5)\n",
    "sns.set(style='whitegrid', font_scale=1.5)\n",
    "sns.residplot(x=y_predict, y=lm_sci_resid, lowess=True, line_kws={'color':'r', 'lw':1.5})\n",
    "ax1.set_title('Residual Plot w.r.t. Training Set', size = 20)\n",
    "ax1.set_xlabel('Fitted Values (Predicted Response)', size = 20)\n",
    "ax1.set_ylabel('Residuals', size = 20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Residual 선이 수평선 0 에 줄곧 붙은 모양이고, scatterplot도 특별한 모양이 없으므로 Residual Plot으로 본 이 모델은 괜찮다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "exit 0\n"
     ]
    }
   ],
   "source": [
    "print('exit 0')"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}