Kaggle titanicで機械学習

titanic.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "_cell_guid": "18cd64c8-c9bc-7fa5-2b40-3542a8601a42"
   },
   "source": [
    "# kaggle - Titanic: Machine Learning from Disaster"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "_cell_guid": "7d7c050a-157f-c1ac-d54f-ef9cdba5606a",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "_cell_guid": "bcb77017-6036-8963-9096-996a569b900a",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# データの読み込み\n",
    "df_train = pd.read_csv('data/train.csv') # トレーニングデータ\n",
    "df_test = pd.read_csv('data/test.csv') # テストデータ"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "_cell_guid": "04fbca24-f4b7-6c05-49f3-df6e68abf145",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 学習モデルは数値しか扱えない。Sexはmale, femailと入っているので数値化する\n",
    "\n",
    "# SexId を追加\n",
    "df_train['SexId'] = df_train['Sex'].map({'male': 1, 'female': 0})\n",
    "df_test['SexId'] = df_test['Sex'].map({'male': 1, 'female': 0})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "_cell_guid": "2e9ae43d-a300-f794-8b3b-cfc42b0f69f7",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# ２つのSeriasを合わせてみる\n",
    "\n",
    "# FamilySize = SibSp + Parch\n",
    "df_train['FamilySize'] = df_train['SibSp'] + df_train['Parch']\n",
    "df_test['FamilySize'] = df_test['SibSp'] + df_test['Parch']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "_cell_guid": "cf34978b-4822-f91e-00e1-b5050c7faa53",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 機械学習は欠損値(NaN)を扱えない。\n",
    "# 欠損値を扱う方法は\n",
    "# * 欠損値のあるデータを取り除く\n",
    "# * 欠損値をある数値で埋める\n",
    "#   * 中央値で埋める\n",
    "#   * 平均値で埋める\n",
    "#   * 0で埋める\n",
    "\n",
    "# Ageの欠損値を中央値で補完する\n",
    "df_train['AgeNull'] = df_train['Age'].isnull()\n",
    "age_median = df_train['Age'].median()\n",
    "df_train['Age'].fillna(age_median, inplace=True)\n",
    "df_test['Age'].fillna(age_median, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "_cell_guid": "2ffc3c77-9591-a35f-5d57-a1345016347a",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "inputs = ['FamilySize', 'SexId', 'Age']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "_cell_guid": "0b0633ad-f59b-a367-361c-747d2737fdb7",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# X_train: 学習用の説明変数\n",
    "# y_train: 学習用の目的変数\n",
    "\n",
    "X_train = df_train[inputs].values.astype('float32')\n",
    "y_train = df_train['Survived'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "_cell_guid": "6b1e99e9-59ee-2734-7b70-4ad6f3bfce5e",
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# X_test: 予測するデータの説明変数\n",
    "X_test = df_test[['SexId', 'Age']].values.astype('float32')\n",
    "\n",
    "# y_test: ないの？y_testを予測するのが機械学習のやりたいこと！"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_cell_guid": "8b44b560-2856-ab91-ecb8-d7b4c422ec46"
   },
   "source": [
    "# 機械学習"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "_cell_guid": "d145c5ce-bb78-569d-ce8a-75c50d5a44b6"
   },
   "source": [
    "### RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# ランダムフォレスト\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "_cell_guid": "b1254bdd-950c-04c4-47ea-c7174e57d668",
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# モデルの構築\n",
    "model = RandomForestClassifier(random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "_cell_guid": "73a74338-4d4c-2d11-16f3-bfa4bc4adb6e",
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "            max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=1,\n",
       "            oob_score=False, random_state=42, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# モデルの学習\n",
    "model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "_cell_guid": "6cfcf4e5-286c-5909-09bb-3a0c6983a23f",
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean accuracy (train): 0.8676\n"
     ]
    }
   ],
   "source": [
    "# トレーニングデータに対する予測精度\n",
    "print(\"mean accuracy (train): {0:.4f}\".format(model.score(X_train, y_train)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "_change_revision": 501,
  "_is_fork": false,
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [default]",
   "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}