sklearn.linear_model.LinearRegression vs statsmodel OLS

gistfile1.txt

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import scipy as sp\n",
    "import statsmodels.api as sm\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.linear_model import LinearRegression\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Data from R ISLR package - write.csv(Boston, \"Boston.csv\", col.names = FALSE)\n",
    "boston_df = pd.read_csv(\"../../r/Boston.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(47.117263854857882,\n",
       " array([ -3.05335819e+09,   3.05335819e+09,   9.31299461e-02,\n",
       "         -3.29341722e+00]))"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fitting medv ~ lstat + I(lstat^2)\n",
    "boston_df[\"lstat^2\"] = boston_df[\"lstat\"] ** 2\n",
    "# fitting medv ~ poly(lstat,4). We already have lstat^2 and lstat from previous\n",
    "boston_df[\"lstat^4\"] = np.power(boston_df[\"lstat\"], 4)\n",
    "boston_df[\"lstat^3\"] = np.power(boston_df[\"lstat\"], 4)\n",
    "X = boston_df[[\"lstat^4\", \"lstat^3\", \"lstat^2\", \"lstat\"]]\n",
    "y = boston_df[\"medv\"]\n",
    "reg7 = LinearRegression()\n",
    "reg7.fit(X, y)\n",
    "(reg7.intercept_, reg7.coef_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# X = boston_df[[\"lstat^4\", \"lstat^3\", \"lstat^2\", \"lstat\"]]\n",
    "X = sm.add_constant(X)\n",
    "# X = boston_df[[1., \"lstat^4\", \"lstat^3\", \"lstat^2\", \"lstat\"]]\n",
    "ols = sm.OLS(y,X).fit()\n",
    "# ols.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "eps = 0.0000000001\n",
    "np.all(np.abs(ols.params.values[1:] - reg7.coef_) < eps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ -1.17513710e-05,  -1.17509020e-05,   9.23027375e-02,\n",
       "        -3.27115207e+00])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ols.params.values[1:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}