ShelterAnimals.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from datetime import date, datetime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train = pd.read_csv('train.csv')\n",
    "test = pd.read_csv('test.csv')"
   ]
  },
  {
   "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>AnimalID</th>\n",
       "      <th>Name</th>\n",
       "      <th>DateTime</th>\n",
       "      <th>OutcomeType</th>\n",
       "      <th>OutcomeSubtype</th>\n",
       "      <th>AnimalType</th>\n",
       "      <th>SexuponOutcome</th>\n",
       "      <th>AgeuponOutcome</th>\n",
       "      <th>Breed</th>\n",
       "      <th>Color</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>A671945</td>\n",
       "      <td>Hambone</td>\n",
       "      <td>2014-02-12 18:22:00</td>\n",
       "      <td>Return_to_owner</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Dog</td>\n",
       "      <td>Neutered Male</td>\n",
       "      <td>1 year</td>\n",
       "      <td>Shetland Sheepdog Mix</td>\n",
       "      <td>Brown/White</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>A656520</td>\n",
       "      <td>Emily</td>\n",
       "      <td>2013-10-13 12:44:00</td>\n",
       "      <td>Euthanasia</td>\n",
       "      <td>Suffering</td>\n",
       "      <td>Cat</td>\n",
       "      <td>Spayed Female</td>\n",
       "      <td>1 year</td>\n",
       "      <td>Domestic Shorthair Mix</td>\n",
       "      <td>Cream Tabby</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  AnimalID     Name             DateTime      OutcomeType OutcomeSubtype  \\\n",
       "0  A671945  Hambone  2014-02-12 18:22:00  Return_to_owner            NaN   \n",
       "1  A656520    Emily  2013-10-13 12:44:00       Euthanasia      Suffering   \n",
       "\n",
       "  AnimalType SexuponOutcome AgeuponOutcome                   Breed  \\\n",
       "0        Dog  Neutered Male         1 year   Shetland Sheepdog Mix   \n",
       "1        Cat  Spayed Female         1 year  Domestic Shorthair Mix   \n",
       "\n",
       "         Color  \n",
       "0  Brown/White  \n",
       "1  Cream Tabby  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>ID</th>\n",
       "      <th>Name</th>\n",
       "      <th>DateTime</th>\n",
       "      <th>AnimalType</th>\n",
       "      <th>SexuponOutcome</th>\n",
       "      <th>AgeuponOutcome</th>\n",
       "      <th>Breed</th>\n",
       "      <th>Color</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>Summer</td>\n",
       "      <td>2015-10-12 12:15:00</td>\n",
       "      <td>Dog</td>\n",
       "      <td>Intact Female</td>\n",
       "      <td>10 months</td>\n",
       "      <td>Labrador Retriever Mix</td>\n",
       "      <td>Red/White</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>Cheyenne</td>\n",
       "      <td>2014-07-26 17:59:00</td>\n",
       "      <td>Dog</td>\n",
       "      <td>Spayed Female</td>\n",
       "      <td>2 years</td>\n",
       "      <td>German Shepherd/Siberian Husky</td>\n",
       "      <td>Black/Tan</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   ID      Name             DateTime AnimalType SexuponOutcome AgeuponOutcome  \\\n",
       "0   1    Summer  2015-10-12 12:15:00        Dog  Intact Female      10 months   \n",
       "1   2  Cheyenne  2014-07-26 17:59:00        Dog  Spayed Female        2 years   \n",
       "\n",
       "                            Breed      Color  \n",
       "0          Labrador Retriever Mix  Red/White  \n",
       "1  German Shepherd/Siberian Husky  Black/Tan  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Some utility functions for assisting in feature creation.\n",
    "#### For the following unknowns are gettting recoded to -999"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# From: http://stackoverflow.com/a/28688724\n",
    "Y = 2000 # dummy leap year to allow input X-02-29 (leap day)\n",
    "seasons = [('winter', (date(Y,  1,  1),  date(Y,  3, 20))),\n",
    "           ('spring', (date(Y,  3, 21),  date(Y,  6, 20))),\n",
    "           ('summer', (date(Y,  6, 21),  date(Y,  9, 22))),\n",
    "           ('autumn', (date(Y,  9, 23),  date(Y, 12, 20))),\n",
    "           ('winter', (date(Y, 12, 21),  date(Y, 12, 31)))]\n",
    "\n",
    "def get_season(now):\n",
    "    if isinstance(now, datetime):\n",
    "        now = now.date()\n",
    "    now = now.replace(year=Y)\n",
    "    return next(season for season, (start, end) in seasons\n",
    "                if start <= now <= end)\n",
    "\n",
    "def get_breed(line):\n",
    "    # mix is the last word\n",
    "    l = line.split('/')\n",
    "    mix = 1 if 'Mix' in line else 0\n",
    "    \n",
    "    if mix:\n",
    "        l = line.split(' ')\n",
    "        # remove \n",
    "        del l[-1]\n",
    "        l = ' '.join(l).split('/')\n",
    "\n",
    "    if len(l) == 2:\n",
    "        # remember We need to return series to add to multiple columns\n",
    "        return pd.Series({'breed_1': l[0], 'breed_2': l[1], 'is_mix': mix})\n",
    "\n",
    "    return pd.Series({'breed_1': l[0], 'breed_2': 'NaN', 'is_mix': mix})\n",
    "\n",
    "def age_years(x):\n",
    "    age = int(x.split(' ')[0])\n",
    "\n",
    "    if 'month' in x:\n",
    "        age /= 12.0\n",
    "\n",
    "    elif 'week' in x:\n",
    "        age /= 52.0\n",
    "\n",
    "    elif 'day' in x:\n",
    "        age /= 365.0\n",
    "\n",
    "    else:\n",
    "        age = float(age)\n",
    "\n",
    "    return age\n",
    "\n",
    "def gender(string):\n",
    "    try:\n",
    "        if 'Female' in string:\n",
    "            return 1\n",
    "\n",
    "        if 'Male' in string:\n",
    "            return 0\n",
    "        \n",
    "        return -999\n",
    "\n",
    "    except:\n",
    "        return -999\n",
    "    \n",
    "def fixed(string):\n",
    "    try:\n",
    "        if 'Spayed' in string or 'Neutered' in string:\n",
    "            return 1\n",
    "\n",
    "        if 'Unknown' in string:\n",
    "            return -999\n",
    "\n",
    "        return 0\n",
    "    except:\n",
    "        return -999"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### We are going to modify all the data test+train together to generate our new features then break them back up for simplicity sake."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# get number of training elements\n",
    "train_num = train.shape[0]\n",
    "\n",
    "#df.rename(columns={'$a': 'a', '$b': 'b'}, inplace=True)\n",
    "train.rename(columns={'AnimalID': 'ID'}, inplace=True)\n",
    "\n",
    "all_data = pd.concat((train, test), axis=0, ignore_index=True)\n",
    "\n",
    "# create a new data frame to store our new features.\n",
    "new_data = pd.DataFrame()\n",
    "\n",
    "#We need an index to work with\n",
    "new_data['ID'] = all_data['ID']\n",
    "\n",
    "# Add the easy stuff to our new dataframe\n",
    "# is it a cat?\n",
    "new_data['is_cat'] = all_data['AnimalType'].map(lambda x: 1 if 'Cat' in x else 0) \n",
    "\n",
    "# color stuff\n",
    "new_data = new_data.join(all_data['Color'].apply(\n",
    "        lambda x: pd.Series({'color_1':x.split('/')[0], 'color_2':x.split('/')[1]}\n",
    "                            if len(x.split('/')) == 2 else {'color_1':x, 'color_2':'NaN'})))\n",
    "\n",
    "# lets convert the date into seasons\n",
    "new_data = new_data.join(\n",
    "    pd.get_dummies(\n",
    "        all_data['DateTime'].map(lambda x: get_season(datetime.strptime(x, '%Y-%m-%d %H:%M:%S')))\n",
    "    ))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# add in the breed information\n",
    "new_data = new_data.join(all_data['Breed'].apply(lambda x: get_breed(x)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>AgeuponOutcome</th>\n",
       "      <th>AnimalType</th>\n",
       "      <th>Breed</th>\n",
       "      <th>Color</th>\n",
       "      <th>DateTime</th>\n",
       "      <th>ID</th>\n",
       "      <th>Name</th>\n",
       "      <th>OutcomeSubtype</th>\n",
       "      <th>OutcomeType</th>\n",
       "      <th>SexuponOutcome</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1 year</td>\n",
       "      <td>Dog</td>\n",
       "      <td>Shetland Sheepdog Mix</td>\n",
       "      <td>Brown/White</td>\n",
       "      <td>2014-02-12 18:22:00</td>\n",
       "      <td>A671945</td>\n",
       "      <td>Hambone</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Return_to_owner</td>\n",
       "      <td>Neutered Male</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1 year</td>\n",
       "      <td>Cat</td>\n",
       "      <td>Domestic Shorthair Mix</td>\n",
       "      <td>Cream Tabby</td>\n",
       "      <td>2013-10-13 12:44:00</td>\n",
       "      <td>A656520</td>\n",
       "      <td>Emily</td>\n",
       "      <td>Suffering</td>\n",
       "      <td>Euthanasia</td>\n",
       "      <td>Spayed Female</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2 years</td>\n",
       "      <td>Dog</td>\n",
       "      <td>Pit Bull Mix</td>\n",
       "      <td>Blue/White</td>\n",
       "      <td>2015-01-31 12:28:00</td>\n",
       "      <td>A686464</td>\n",
       "      <td>Pearce</td>\n",
       "      <td>Foster</td>\n",
       "      <td>Adoption</td>\n",
       "      <td>Neutered Male</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3 weeks</td>\n",
       "      <td>Cat</td>\n",
       "      <td>Domestic Shorthair Mix</td>\n",
       "      <td>Blue Cream</td>\n",
       "      <td>2014-07-11 19:09:00</td>\n",
       "      <td>A683430</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Partner</td>\n",
       "      <td>Transfer</td>\n",
       "      <td>Intact Male</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2 years</td>\n",
       "      <td>Dog</td>\n",
       "      <td>Lhasa Apso/Miniature Poodle</td>\n",
       "      <td>Tan</td>\n",
       "      <td>2013-11-15 12:52:00</td>\n",
       "      <td>A667013</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Partner</td>\n",
       "      <td>Transfer</td>\n",
       "      <td>Neutered Male</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  AgeuponOutcome AnimalType                        Breed        Color  \\\n",
       "0         1 year        Dog        Shetland Sheepdog Mix  Brown/White   \n",
       "1         1 year        Cat       Domestic Shorthair Mix  Cream Tabby   \n",
       "2        2 years        Dog                 Pit Bull Mix   Blue/White   \n",
       "3        3 weeks        Cat       Domestic Shorthair Mix   Blue Cream   \n",
       "4        2 years        Dog  Lhasa Apso/Miniature Poodle          Tan   \n",
       "\n",
       "              DateTime       ID     Name OutcomeSubtype      OutcomeType  \\\n",
       "0  2014-02-12 18:22:00  A671945  Hambone            NaN  Return_to_owner   \n",
       "1  2013-10-13 12:44:00  A656520    Emily      Suffering       Euthanasia   \n",
       "2  2015-01-31 12:28:00  A686464   Pearce         Foster         Adoption   \n",
       "3  2014-07-11 19:09:00  A683430      NaN        Partner         Transfer   \n",
       "4  2013-11-15 12:52:00  A667013      NaN        Partner         Transfer   \n",
       "\n",
       "  SexuponOutcome  \n",
       "0  Neutered Male  \n",
       "1  Spayed Female  \n",
       "2  Neutered Male  \n",
       "3    Intact Male  \n",
       "4  Neutered Male  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_data.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'float' object has no attribute 'split'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-8-90c6981b7cd9>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mnew_data\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'age_years'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mall_data\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'AgeuponOutcome'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mage_years\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m//anaconda/lib/python2.7/site-packages/pandas/core/series.pyc\u001b[0m in \u001b[0;36mmap\u001b[0;34m(self, arg, na_action)\u001b[0m\n\u001b[1;32m   2043\u001b[0m                                      index=self.index).__finalize__(self)\n\u001b[1;32m   2044\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2045\u001b[0;31m             \u001b[0mmapped\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmap_f\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2046\u001b[0m             return self._constructor(mapped,\n\u001b[1;32m   2047\u001b[0m                                      index=self.index).__finalize__(self)\n",
      "\u001b[0;32mpandas/src/inference.pyx\u001b[0m in \u001b[0;36mpandas.lib.map_infer (pandas/lib.c:62187)\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32m<ipython-input-5-27d31c4dc7c1>\u001b[0m in \u001b[0;36mage_years\u001b[0;34m(x)\u001b[0m\n\u001b[1;32m     32\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     33\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mage_years\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 34\u001b[0;31m     \u001b[0mage\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msplit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m' '\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     35\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     36\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0;34m'month'\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'float' object has no attribute 'split'"
     ]
    }
   ],
   "source": [
    "new_data['age_years'] = all_data['AgeuponOutcome'].map(age_years)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# what is the gender, I am coding all unknowns as -999\n",
    "new_data['is_female'] = all_data['SexuponOutcome'].map(gender)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# are they fixed\n",
    "new_data['is_fixed'] = all_data['SexuponOutcome'].map(fixed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## We are going to recode breed and color so we have a sparse array.\n",
    "### I am also recoding age, but it's probably not the best approach because it is using the quartiles of the two data sets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "\n",
    "new_data = new_data.join(\n",
    "    pd.get_dummies(\n",
    "        new_data['breed_1'], prefix='breed')\n",
    "    )\n",
    "\n",
    "new_data = new_data.join(\n",
    "    pd.get_dummies(\n",
    "        new_data['color_1'], prefix='color')\n",
    "    )\n",
    "\n",
    "# This is probably an inappropriate way to dummy code the ages \n",
    "# but just to get something together for testing\n",
    "new_data = new_data.join(\n",
    "    pd.get_dummies(\n",
    "        pd.qcut(new_data['age_years'], 4, labels=[\"age_1\",\"age_2\",\"age_3\", \"age_4\"]))\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "new_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# We want to drop the original non-binary columns now. \n",
    "cols_to_drop = ['color_1',\n",
    "                'color_2',\n",
    "                'breed_1',\n",
    "                'breed_2',\n",
    "                'age_years',\n",
    "                'ID']\n",
    "new_data = new_data.drop(cols_to_drop, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Now we have a binary feature matrix."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "new_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Lets break things back up into our test and train data sets.\n",
    "X_train_all = new_data.iloc[:train_num]\n",
    "y_train_all = all_data['OutcomeType'][:train_num]\n",
    "\n",
    "X_test_all = new_data.iloc[train_num:]\n",
    "ids_test = all_data['ID'][train_num:].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X_train_all.shape, y_train_all.shape, X_test_all.shape, ids_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Just to get things going..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.cross_validation import train_test_split\n",
    "from sklearn.metrics import log_loss\n",
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X_train_all, y_train_all, test_size=0.30, random_state=23)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "lr = LogisticRegression(random_state=23, max_iter=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "y_pred = lr.fit(X_train, y_train).predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "print np.mean(y_pred == y_test.values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Not particularly good, but whatever. To create a submission file I think we need the "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Not particularly good, but whatever. To create a submission file I think we need the \n",
    "# robability for each class. First we retrain on entire dataset then classify the test data.\n",
    "y_pred_sub = lr.fit(X_train_all, y_train_all).predict_proba(X_test_all)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Prepare the submission file..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Prepare the submission file\n",
    "sub = pd.DataFrame(y_pred_sub, columns=['Adoption', 'Died', 'Euthanasia', 'Return_to_owner', 'Transfer'])\n",
    "sub.insert(0, 'ID', ids_test.astype(int))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "sub.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "sub.to_csv(\"submission.csv\", index=False)"
   ]
  },
  {
   "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": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}