ML course - linear methods asessment

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Нормализация признаков\n",
    "\n",
    "## Задание курса Machine learning - неделя 2 - линейные методы"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.linear_model import Perceptron\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Чтение данных"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "train = pd.read_csv('perceptron-train.csv', header=None)\n",
    "test = pd.read_csv('perceptron-test.csv', header=None)\n",
    "X_train = train.iloc[:,1:].values\n",
    "Y_train = train.iloc[:, 0].values\n",
    "X_test = test.iloc[:,1:].values\n",
    "Y_test = test.iloc[:, 0].values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  Обучение перцептрона без нормализации"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Точность без нормализации - 0.360\n"
     ]
    }
   ],
   "source": [
    "clf = Perceptron(random_state=241)\n",
    "clf.fit(X_train, Y_train)\n",
    "score1 = clf.score(X_test, Y_test)\n",
    "print 'Точность без нормализации - {:.3f}'.format(score1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Нормализация и обучение нового перцептрона"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "scaler = StandardScaler()\n",
    "X_train_s = scaler.fit_transform(X_train)\n",
    "X_test_s = scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Точность после нормализации - 0.925\n",
      "Разность в результатах - 0.565\n"
     ]
    }
   ],
   "source": [
    "clf.fit(X_train_s, Y_train)\n",
    "score2 = clf.score(X_test_s, Y_test)\n",
    "print 'Точность после нормализации - {:.3f}'.format(score2)\n",
    "diff = score2 - score1\n",
    "print 'Разность в результатах - {:.3f}'.format(diff)\n",
    "with open('ans.txt', 'w') as f:\n",
    "    f.write('{:.3f}'.format(diff))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Повторение эксперимента, если задать столбец ответов как Boolean - повышает точность работы"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Y_train = train.iloc[:, 0].values.astype(np.bool_)\n",
    "Y_test = test.iloc[:, 0].values.astype(np.bool_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Точность без нормализации - 0.675\n"
     ]
    }
   ],
   "source": [
    "clf = Perceptron(random_state=241)\n",
    "clf.fit(X_train, Y_train)\n",
    "score1 = clf.score(X_test, Y_test)\n",
    "print 'Точность без нормализации - {:.3f}'.format(score1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Точность после нормализации - 0.970\n",
      "Разность в результатах - 0.295\n"
     ]
    }
   ],
   "source": [
    "clf.fit(X_train_s, Y_train)\n",
    "score2 = clf.score(X_test_s, Y_test)\n",
    "print 'Точность после нормализации - {:.3f}'.format(score2)\n",
    "diff = score2 - score1\n",
    "print 'Разность в результатах - {:.3f}'.format(diff)\n",
    "with open('ans.txt', 'w') as f:\n",
    "    f.write('{:.3f}'.format(diff))"
   ]
  }
 ],
 "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": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}