Introduction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Практическое руководство по анализу данных\n",
    "\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Рабочее окружение: Jupyter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Ноутбуки (тетрадки)\n",
    "\n",
    "- Разметка\n",
    "- Возможность вывода графиков, изображений и тд."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Редактирование python-файлов\n",
    "\n",
    "- Как создавать файлы\n",
    "- Идея: выносить сложные куски кода в python-модули, в ноутбуке оставлять эксперименты"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Работа в консоли\n",
    "\n",
    "- Вызов команд консоли из ноутбука: `!ls -l`\n",
    "- Использование терминала"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "wget: missing URL\r\n",
      "Usage: wget [OPTION]... [URL]...\r\n",
      "\r\n",
      "Try `wget --help' for more options.\r\n"
     ]
    }
   ],
   "source": [
    "!wget"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dockerfile           adult.data           environment.yml\r\n",
      "Dockerfile.datanight advanced.ipynb       \u001b[34mscripts\u001b[m\u001b[m\r\n",
      "Introduction.ipynb   baseline.ipynb       submit.py\r\n",
      "README.md            datanight.py         template.ipynb\r\n"
     ]
    }
   ],
   "source": [
    "!ls"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tar: Option f requires an argument\r\n",
      "Usage:\r\n",
      "  List:    tar -tf <archive-filename>\r\n",
      "  Extract: tar -xf <archive-filename>\r\n",
      "  Create:  tar -cf <archive-filename> [filenames...]\r\n",
      "  Help:    tar --help\r\n"
     ]
    }
   ],
   "source": [
    "!tar xvf "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "UnZip 5.52 of 28 February 2005, by Info-ZIP.  Maintained by C. Spieler.  Send\r\n",
      "bug reports using http://www.info-zip.org/zip-bug.html; see README for details.\r\n",
      "\r\n",
      "Usage: unzip [-Z] [-opts[modifiers]] file[.zip] [list] [-x xlist] [-d exdir]\r\n",
      "  Default action is to extract files in list, except those in xlist, to exdir;\r\n",
      "  file[.zip] may be a wildcard.  -Z => ZipInfo mode (\"unzip -Z\" for usage).\r\n",
      "\r\n",
      "  -p  extract files to pipe, no messages     -l  list files (short format)\r\n",
      "  -f  freshen existing files, create none    -t  test compressed archive data\r\n",
      "  -u  update files, create if necessary      -z  display archive comment\r\n",
      "  -x  exclude files that follow (in xlist)   -d  extract files into exdir\r\n",
      "\r\n",
      "modifiers:                                   -q  quiet mode (-qq => quieter)\r\n",
      "  -n  never overwrite existing files         -a  auto-convert any text files\r\n",
      "  -o  overwrite files WITHOUT prompting      -aa treat ALL files as text\r\n",
      "  -j  junk paths (do not make directories)   -v  be verbose/print version info\r\n",
      "  -C  match filenames case-insensitively     -L  make (some) names lowercase\r\n",
      "  -X  restore UID/GID info                   -V  retain VMS version numbers\r\n",
      "  -K  keep setuid/setgid/tacky permissions   -M  pipe through \"more\" pager\r\n",
      "Examples (see unzip.txt for more info):\r\n",
      "  unzip data1 -x joe   => extract all files except joe from zipfile data1.zip\r\n",
      "  unzip -p foo | more  => send contents of foo.zip via pipe into program more\r\n",
      "  unzip -fo foo ReadMe => quietly replace existing ReadMe if archive file newer\r\n"
     ]
    }
   ],
   "source": [
    "!unzip"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Получение и работа с данными"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Табличные данные"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "pandas.read_csv('http://')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "pandas.read_csv('')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Таблицы со станиц интернета"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tables = pandas.read_html(\n",
    "    'https://en.wikipedia.org/w/index.php?search=Dow+Jones+Industrial+Average&title=Special%3ASearch&go=Go', \n",
    "    attrs={'class': 'wikitable'},  # фильтруем только таблицы с данными\n",
    "    header=0, # делаем первую строчку — заголовком\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "tables[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Использование HTTP API\n",
    "\n",
    "Пример API:\n",
    " - Яндекс Карты"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import requests\n",
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "resp = requests.get(\n",
    "    'https://geocode-maps.yandex.ru/1.x/',\n",
    "    params={'format': 'json', 'geocode': 'Масква Тверская улца дом 7'},\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "found = resp.json()['response']['GeoObjectCollection']['featureMember'][0]['GeoObject']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{\n",
      "  \"metaDataProperty\": {\n",
      "    \"GeocoderMetaData\": {\n",
      "      \"text\": \"Россия, Москва, Тверская улица, 7\", \n",
      "      \"kind\": \"house\", \n",
      "      \"AddressDetails\": {\n",
      "        \"Country\": {\n",
      "          \"CountryName\": \"Россия\", \n",
      "          \"AdministrativeArea\": {\n",
      "            \"AdministrativeAreaName\": \"Москва\", \n",
      "            \"Locality\": {\n",
      "              \"Thoroughfare\": {\n",
      "                \"Premise\": {\n",
      "                  \"PremiseNumber\": \"7\"\n",
      "                }, \n",
      "                \"ThoroughfareName\": \"Тверская улица\"\n",
      "              }, \n",
      "              \"LocalityName\": \"Москва\"\n",
      "            }\n",
      "          }, \n",
      "          \"AddressLine\": \"Москва, Тверская улица, 7\", \n",
      "          \"CountryNameCode\": \"RU\"\n",
      "        }\n",
      "      }, \n",
      "      \"precision\": \"exact\"\n",
      "    }\n",
      "  }, \n",
      "  \"boundedBy\": {\n",
      "    \"Envelope\": {\n",
      "      \"upperCorner\": \"37.619234 55.762601\", \n",
      "      \"lowerCorner\": \"37.602777 55.753321\"\n",
      "    }\n",
      "  }, \n",
      "  \"Point\": {\n",
      "    \"pos\": \"37.611006 55.757962\"\n",
      "  }, \n",
      "  \"description\": \"Москва, Россия\", \n",
      "  \"name\": \"Тверская улица, 7\"\n",
      "}\n"
     ]
    }
   ],
   "source": [
    "print json.dumps(found, indent=2, encoding='utf8', ensure_ascii=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Извлечение данных со страниц\n",
    "\n",
    "- [Scrapy](http://scrapy.org/)\n",
    "- [Пример работы с селекторами](http://doc.scrapy.org/en/1.0/topics/selectors.html)\n",
    "- [Документация по css селекторам](http://www.w3schools.com/cssref/css_selectors.asp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "resp = requests.get('https://www.hse.ru/staff/arjanstev')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import scrapy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "page = scrapy.Selector(text=resp.content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "u'<dl class=\"main-list large main-list-language-knowledge-level\"><dt class=\"b\">\\u0412\\u043b\\u0430\\u0434\\u0435\\u043d\\u0438\\u0435 \\u044f\\u0437\\u044b\\u043a\\u0430\\u043c\\u0438</dt><dd>\\u0430\\u043d\\u0433\\u043b\\u0438\\u0439\\u0441\\u043a\\u0438\\u0439</dd><dd>\\u0443\\u043a\\u0440\\u0430\\u0438\\u043d\\u0441\\u043a\\u0438\\u0439</dd><dd>\\u043d\\u0435\\u043c\\u0435\\u0446\\u043a\\u0438\\u0439</dd><dd>\\u0444\\u0440\\u0430\\u043d\\u0446\\u0443\\u0437\\u0441\\u043a\\u0438\\u0439</dd></dl>'"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "page.css('dl.main-list-language-knowledge-level').extract_first()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "английский\n",
      "украинский\n",
      "немецкий\n",
      "французский\n"
     ]
    }
   ],
   "source": [
    "langs = page.css('dl.main-list-language-knowledge-level dd ::text').extract()\n",
    "for lang in langs:\n",
    "    print lang"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Таблица из набора JSON-записей"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "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>x</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   x  y\n",
       "0  1  a\n",
       "1  2  b\n",
       "2  3  a\n",
       "3  4  b"
      ]
     },
     "execution_count": 168,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pandas.DataFrame({\n",
    "        'x': [1, 2, 3, 4],\n",
    "        'y': ['a', 'b', 'a', 'b']\n",
    "    })"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "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>x</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>a</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>b</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   x  y\n",
       "0  1  a\n",
       "1  2  b"
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pandas.DataFrame.from_records([\n",
    "        {'x': 1, 'y': 'a'},\n",
    "        {'x': 2, 'y': 'b'}\n",
    "    ])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Машинное обучение"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Рассмотрим выборку [Census Income Data Set](https://archive.ics.uci.edu/ml/machine-learning-databases/adult/)\n",
    "\n",
    "Решаем следующую задачу - по социальным признакам человека необходимо определить зарабатывает ли он более 50к или менее."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Предобработка данных, извлечение признаков"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Качаем выборку"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2016-04-08 13:11:51--  https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data\n",
      "Resolving archive.ics.uci.edu... 128.195.10.249\n",
      "Connecting to archive.ics.uci.edu|128.195.10.249|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 3974305 (3.8M) [text/plain]\n",
      "Saving to: 'adult.data'\n",
      "\n",
      "adult.data          100%[=====================>]   3.79M  1.49MB/s   in 2.5s   \n",
      "\n",
      "2016-04-08 13:11:54 (1.49 MB/s) - 'adult.data' saved [3974305/3974305]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Посмотрим как выглядит начало файла с данными"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "39, State-gov, 77516, Bachelors, 13, Never-married, Adm-clerical, Not-in-family, White, Male, 2174, 0, 40, United-States, <=50K\r\n",
      "50, Self-emp-not-inc, 83311, Bachelors, 13, Married-civ-spouse, Exec-managerial, Husband, White, Male, 0, 0, 13, United-States, <=50K\r\n",
      "38, Private, 215646, HS-grad, 9, Divorced, Handlers-cleaners, Not-in-family, White, Male, 0, 0, 40, United-States, <=50K\r\n",
      "53, Private, 234721, 11th, 7, Married-civ-spouse, Handlers-cleaners, Husband, Black, Male, 0, 0, 40, United-States, <=50K\r\n",
      "28, Private, 338409, Bachelors, 13, Married-civ-spouse, Prof-specialty, Wife, Black, Female, 0, 0, 40, Cuba, <=50K\r\n",
      "37, Private, 284582, Masters, 14, Married-civ-spouse, Exec-managerial, Wife, White, Female, 0, 0, 40, United-States, <=50K\r\n",
      "49, Private, 160187, 9th, 5, Married-spouse-absent, Other-service, Not-in-family, Black, Female, 0, 0, 16, Jamaica, <=50K\r\n",
      "52, Self-emp-not-inc, 209642, HS-grad, 9, Married-civ-spouse, Exec-managerial, Husband, White, Male, 0, 0, 45, United-States, >50K\r\n",
      "31, Private, 45781, Masters, 14, Never-married, Prof-specialty, Not-in-family, White, Female, 14084, 0, 50, United-States, >50K\r\n",
      "42, Private, 159449, Bachelors, 13, Married-civ-spouse, Exec-managerial, Husband, White, Male, 5178, 0, 40, United-States, >50K\r\n"
     ]
    }
   ],
   "source": [
    "!head adult.data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Импортируем необходимые библиотеки"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas\n",
    "import numpy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Загружаем файл и делаем из него pandas.DataFrame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(32561, 15)\n"
     ]
    },
    {
     "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",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>39</td>\n",
       "      <td>State-gov</td>\n",
       "      <td>77516</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>13</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Adm-clerical</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>2174</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>50</td>\n",
       "      <td>Self-emp-not-inc</td>\n",
       "      <td>83311</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>13</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>215646</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>9</td>\n",
       "      <td>Divorced</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>53</td>\n",
       "      <td>Private</td>\n",
       "      <td>234721</td>\n",
       "      <td>11th</td>\n",
       "      <td>7</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Handlers-cleaners</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>28</td>\n",
       "      <td>Private</td>\n",
       "      <td>338409</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>13</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Prof-specialty</td>\n",
       "      <td>Wife</td>\n",
       "      <td>Black</td>\n",
       "      <td>Female</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>Cuba</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0                  1       2           3   4                    5   \\\n",
       "0  39          State-gov   77516   Bachelors  13        Never-married   \n",
       "1  50   Self-emp-not-inc   83311   Bachelors  13   Married-civ-spouse   \n",
       "2  38            Private  215646     HS-grad   9             Divorced   \n",
       "3  53            Private  234721        11th   7   Married-civ-spouse   \n",
       "4  28            Private  338409   Bachelors  13   Married-civ-spouse   \n",
       "\n",
       "                   6               7       8        9     10  11  12  \\\n",
       "0        Adm-clerical   Not-in-family   White     Male  2174   0  40   \n",
       "1     Exec-managerial         Husband   White     Male     0   0  13   \n",
       "2   Handlers-cleaners   Not-in-family   White     Male     0   0  40   \n",
       "3   Handlers-cleaners         Husband   Black     Male     0   0  40   \n",
       "4      Prof-specialty            Wife   Black   Female     0   0  40   \n",
       "\n",
       "               13      14  \n",
       "0   United-States   <=50K  \n",
       "1   United-States   <=50K  \n",
       "2   United-States   <=50K  \n",
       "3   United-States   <=50K  \n",
       "4            Cuba   <=50K  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pandas.read_csv('adult.data', header=None)\n",
    "print data.shape\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Выберем непрерывные (числовые) признаки и дискретные (категориальные) признаки."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "continious_features = [0,2,4,10,11,12]\n",
    "discreete_features = [1,3,5,6,7,8,9,13]\n",
    "target = [14]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Создадим матрицу, где строчкой будет объект, а столбцом признак. Сначала составим матрицу из непрерывных признаков."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(32561, 6)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[    39,  77516,     13,   2174,      0,     40],\n",
       "       [    50,  83311,     13,      0,      0,     13],\n",
       "       [    38, 215646,      9,      0,      0,     40],\n",
       "       ..., \n",
       "       [    58, 151910,      9,      0,      0,     40],\n",
       "       [    22, 201490,      9,      0,      0,     20],\n",
       "       [    52, 287927,      9,  15024,      0,     40]])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = data.iloc[:,continious_features].values\n",
    "print X.shape\n",
    "X"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Обработка категориальных признаков"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import LabelBinarizer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Для каждого категориального признака последовательно бинаризуем его и добавим получившиеся признаки к уже существующим"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(32561, 107)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[    39,  77516,     13, ...,      1,      0,      0],\n",
       "       [    50,  83311,     13, ...,      1,      0,      0],\n",
       "       [    38, 215646,      9, ...,      1,      0,      0],\n",
       "       ..., \n",
       "       [    58, 151910,      9, ...,      1,      0,      0],\n",
       "       [    22, 201490,      9, ...,      1,      0,      0],\n",
       "       [    52, 287927,      9, ...,      1,      0,      0]])"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lb = LabelBinarizer()\n",
    "for f in discreete_features:\n",
    "    X = numpy.hstack([X,lb.fit_transform(data.iloc[:,f])])\n",
    "\n",
    "print X.shape\n",
    "X"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Также бинаризуем таргет"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(32561, 1)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[0],\n",
       "       [0],\n",
       "       [0],\n",
       "       ..., \n",
       "       [0],\n",
       "       [0],\n",
       "       [1]])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y = lb.fit_transform(data.iloc[:,target])\n",
    "print Y.shape\n",
    "Y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Наконец разбиваем выборку на обучающую часть и тестовую"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "ids = numpy.arange(X.shape[0])\n",
    "numpy.random.shuffle(ids)\n",
    "\n",
    "train_len = int(len(ids) * 0.8)\n",
    "\n",
    "X_train = X[ids[:train_len],:]\n",
    "Y_train = Y[ids[:train_len],:]\n",
    "X_test = X[ids[train_len:],:]\n",
    "Y_test = Y[ids[train_len:],:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Построим гистограммы признаков"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAAHECAYAAAD2y2l8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13627dd90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize(8,6)\n",
    "data[discreete_features[3]].value_counts().plot(kind='bar')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Теперь построим гистограммы для непрерывных признаков, при этом разным цветом - разные таргеты\n",
    "\n",
    "Красным - где таргет 0, синим - таргет 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABJEAAAJZCAYAAAAZJaRRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x134313bd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize(20,10)\n",
    "plt.figure(1)\n",
    "for i in range(len(continious_features)):\n",
    "    plt.subplot(231 + i)\n",
    "    plt.hist(X[np.where(Y == 1)[0], continious_features[i]], bins = 100, alpha = 0.3, color = 'b', normed=True)\n",
    "    plt.hist(X[np.where(Y == 0)[0], continious_features[i]], bins = 100, alpha = 0.3, color = 'r', normed=True)\n",
    "    plt.title(continious_features[i])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Модели машинного обучения\n",
    "\n",
    "К получившейся выборке применим различные модели Машинного Обучения"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Линейные модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "          intercept_scaling=1, max_iter=100, multi_class='ovr', n_jobs=1,\n",
       "          penalty='l2', random_state=None, solver='liblinear', tol=0.0001,\n",
       "          verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LogisticRegression()\n",
    "lr.fit(X_train, Y_train.flatten())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.36644447,  0.63355553],\n",
       "       [ 0.78502011,  0.21497989],\n",
       "       [ 0.74187257,  0.25812743],\n",
       "       ..., \n",
       "       [ 0.6856135 ,  0.3143865 ],\n",
       "       [ 0.77422698,  0.22577302],\n",
       "       [ 0.77186307,  0.22813693]])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr_preds = lr.predict_proba(X_test)\n",
    "lr_preds"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Деревья, композиции, бустинг"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "            max_depth=5, 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=200, n_jobs=1,\n",
       "            oob_score=False, random_state=None, verbose=0,\n",
       "            warm_start=False)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf = RandomForestClassifier(n_estimators=200, max_depth=5)\n",
    "rf.fit(X_train, Y_train.flatten())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.62201827,  0.37798173],\n",
       "       [ 0.66189608,  0.33810392],\n",
       "       [ 0.72693107,  0.27306893],\n",
       "       ..., \n",
       "       [ 0.72158769,  0.27841231],\n",
       "       [ 0.46696849,  0.53303151],\n",
       "       [ 0.9691158 ,  0.0308842 ]])"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_preds = rf.predict_proba(X_test)\n",
    "rf_preds"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Gradient Boosting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import GradientBoostingClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingClassifier(init=None, learning_rate=0.1, loss='deviance',\n",
       "              max_depth=5, max_features=None, max_leaf_nodes=None,\n",
       "              min_samples_leaf=1, min_samples_split=2,\n",
       "              min_weight_fraction_leaf=0.0, n_estimators=200,\n",
       "              presort='auto', random_state=None, subsample=1.0, verbose=0,\n",
       "              warm_start=False)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bst = GradientBoostingClassifier(n_estimators=200, max_depth=5)\n",
    "bst.fit(X_train, Y_train.flatten())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.0317485 ,  0.9682515 ],\n",
       "       [ 0.34183345,  0.65816655],\n",
       "       [ 0.91602617,  0.08397383],\n",
       "       ..., \n",
       "       [ 0.67166543,  0.32833457],\n",
       "       [ 0.25647266,  0.74352734],\n",
       "       [ 0.99764185,  0.00235815]])"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bst_preds = bst.predict_proba(X_test)\n",
    "bst_preds"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Оценка качества моделей"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Метрики"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_auc_score, roc_curve, accuracy_score\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Качество Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC 0.622973940921\n",
      "accuracy 0.799477967143\n"
     ]
    }
   ],
   "source": [
    "print 'AUC', roc_auc_score(Y_test, lr_preds[:,1])\n",
    "print 'accuracy', accuracy_score(Y_test, lr.predict(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgAAAAIGCAYAAAAiITfnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x134d98e50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr_lr, tpr_lr, thresholds = roc_curve(Y_test, lr_preds[:,1])\n",
    "\n",
    "figsize(8,8)\n",
    "plt.plot(np.arange(0.,1.,1e-5),np.arange(0.,1.,1e-5), c='black')\n",
    "plt.plot(fpr_lr,tpr_lr, c='b')\n",
    "plt.title('ROC curve', size=20)\n",
    "plt.xlabel('False Positive Rate', size=20)\n",
    "plt.ylabel('True Positive Rate', size=20)\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Качество Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC 0.901276750053\n",
      "accuracy 0.839091048672\n"
     ]
    }
   ],
   "source": [
    "print 'AUC', roc_auc_score(Y_test, rf_preds[:,1])\n",
    "print 'accuracy', accuracy_score(Y_test, rf.predict(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgAAAAIGCAYAAAAiITfnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3Xm8XfP1//HXEgkiJF8zUWKIoZR8DTGTGKNaMRatuYNO\nplYNv1alhqJfKvX1bQ2lVBUtKuahuIaYS1ASQSSIIRIJzSjD+v3xOVdObu5w9r3ns6fzfj4e93Ht\nc/bdd93V06zP3p+1P9vcHREREWksS2QdgIiIiKRPAwAREZEGpAGAiIhIA9IAQEREpAFpACAiItKA\nNAAQERFpQBoAiIiINCANAEQyYmYLWnzNM7MpZvaImR1Vw8/vbmY3m9kEM5tlZlPN7Fkz+6WZ9eng\nZ83MDjKzW83sncrPTzez18zsCjPbvn5/qYjkkWkhIJFsmNkCwIFhgAHdgfWB/YEewP+5+/Gt/FwP\n4GrgW8BM4F5gLNAL2BXYBJgMHOjuj7fy86sCtwLbA58BDwJvVWJYv3KM5YHj3f33dfuDRSRXNAAQ\nyUjzAMDdu7V4fTuguXCv7+7jW7x/DXA08Dywn7u/3+L9HwKXAjOAge7+etV7ywBPAV8BbgR+5O6f\ntvj5nsBPgXnufn4X/0wRySlNAYjkjLs/BYwhnJFvWf2eme1AKP6fAF9rWfwrP/974H+A5QgDgWo/\nATYDnnD3w1sW/8rPz3T3c4CLao3ZzLauTEe8Z2azzex9M7vfzA6u2meXylTHL9s4xngzG9fitaMq\nP3OkmQ2pTI9MM7P5ZrZG5fu/2onr3srPf7nF69uY2S1m9oGZzalMg1xuZqvX+jeLFJ0GACL5NqfF\n9vcI0wZXuvukdn7uwsrP7m5ma1e9/t3Kz5/T0S9297m1BGhm3wWeBPYFRhIGDncBKwM/qOUYzb+y\nndcPBu4kTFn8AbipMvh5EBhgZpu0EtdqwO7Ac+7+WtXrxwJPAHsBDwOXAM8B3waeN7M1E8QsUlhL\nZh2AiCzKzHYGNgJmA8+0eHuHyveH2juGu0+rnBlvV/mZCZXCthYwF3isTrFuDPwf8Cmwo7uPafH+\nGvX4NcDewN7u/mCL964F9gSOAk5t8d4RhJOc66ri6U8YQIwDdnH3D6veG0wYUPwOOLAOcYvkmgYA\nIhkzs7Mq/9kd6A/sB8wHfuDuH7fYvfkS9bs1HPpdwgCguQg3/+wUd/+88xEv4odAN+DslsUfoLUp\nik66vZXiD3A7YfDxLTM7zRdtajqKMNi5qUW8SwInVRf/SqyPmNkdwNfNbFl3n1Gn2EVySQMAkey1\nnBNfABzh7je1tnPObFP5fl/k3/Ncay+6+2wz+xvwHcIl/fsAzGwL4MvAre7+SdWPbFv5PsjMBrZy\nyFUIA5oNgBfrFLtILmkAIJKx5rsAKh362wHXANea2Qfu/miL3T8E+gFfItz6154vVb43n4V/UPm+\nopn1qNNVgOb1BibW4Vjt+bCd964l9DYcxcKByNGE3oHrWuy7YuX7Ke0czwm3VIqUmpoARXLC3We5\n+8PA1wlnodeZ2dItdnui8n339o5VWQio+Q6CkZXjvwe8Qxj471ynsKdVvvetYd8Fle9tnXi0tXiR\n03aDYPNdE28A+5rZ8ma2JHAoYS2Ee1vs3nzXw/Lu3q2NryVbWz9BpGw0ABDJGXd/BbiKcAZ/cou3\n/0hoivuOma3czmF+BiwFPOjuE6pev7Ly87/oKI7KgkMdebryfe8a9p1a+f6llm+Y2fpA7xqO0Zbr\ngKWBQ4B9gJWAG9x9fov9muOt1wBIpLA0ABDJp3OBz4FTzOyLwlg5M72ecCn7bjNb7MzbzL5P6Ij/\nDDipxduXAC8BO5nZ9dXHrvr5Zc1sGGExoI78gdCweGbljoCWx6qOb0wlpqFmtlLVPkuz+HoFSf2Z\ncJXgyMpXa5f/AS4D5gGXVO4IaBlvdzPbsYuxiBSCegBEcsjd3zezy4ETgdOA/1f19vcIUwSHAa+b\n2b2ES+DLAoOBTYGPCUsBj2lx3FlmthdwC/BNQsd7y6WAdyMsIvTjGuIcXVl58A/Ai2Y2ohLLisDW\nhEvuu1X2nWdmvyNcfRhlZv8g/Bu0B6GHoK07BqyGON4zs0cqv2se8Iq7v9TKfq9X1gG4GnjVzO4j\n9FJ0J9wiuRMwidBAKFJqWgpYJCOVpYAXuHurA3EzW4Vwv/oCYL2WtwSa2e6E5rftCIvuzAbeBO4A\n/tfdp9EOMzuQMIgYSLhkvoDQI/A4cI27t1yDoL1jbUNorNuJMJc/GXgZ+KO739Zi31MrcX+J0Nx3\nI/ArYDQhH+tV7XsUoSnyGHf/cwcxfIuFVwJOcffh7ey7CeEKx2BgNcKyye8T+iVudvemWv92kaLK\nfABgZlcDXwM+cvfN2tjnUsIc4wzgaHcflWKIIiIipZOHHoA/Ee7fbZWZ7U04++kPHAdcnlZgIiIi\nZZX5AMDdn2Bhd3BrhhIu61G5JNm78jhTERER6aTMBwA16Muiy55OpLZ7jkVERKQNpboLwMzU0Sgi\nIg3H3Tu8W6alIgwAJrLowiFr0s6yo1k3NZbdsGHDGDZsWNZhlJ7yHF+9cvzhhzBrFtx7LywR6Zrq\nJ5/AP/8Jyy0HM2eG/+7RwTJNn38Oa60F664bJ6aOfPwxrLbaMPbYY1ib+/TtC1/5yuKvL7EEfPnL\n0K1bvPiKrqmpiYMPPpi///3vDB48uFPHyMsAwGj7Xt87gB8BN5vZtsA0d/8otchkEePHj886hIag\nPMf38svjua+VRxhNnAgvvghLdvCv4/PPw2uvwdSpsNpqMGMGfPObcWJdsCAUxD32CNvDhsFWW7X/\nM2YdDxJiO/ro8Zx2WrYxlFF18R80aFCnj5P5AMDM/goMIjyg5B3gLKAH4O5+pbvfY2ZfNbM3CbcB\nHpNdtCKSlalTYdKkhdvucN99odA1mzAhFGVrcTrRWqEHeP11+FKLhYnnzIE+fWCXXdqPZ+214ec/\nhwEDYPXV299XpF7qVfwhB+sA1JOZeZn+njxqamrq8odOOtaIef688mxCd2hqgqefhuuvhylTQkEe\nPx6WXTZcNgaYNy+8d9RRix7nv/8bVm1xn9DSS4eCXj0waGpqYvDgQZH+GoHG/BzH1FbxN7NO9QBo\nACAi0U2cGIp1tTlzYORIeOABeO45mDwZuneHuXPD+3vtFc6yf/SjMPfdrVs4W295di/SCNo789cA\nAA0A0qARfTrynmd3GD0aLr4YHnwwNHytsELb+7//fijmyy+/6OsbbxzO6LfaCnbeGdZcM27c1fKe\n4zJQjuujo8v+nR0AZN4DICLFcd99cPzxofN9+nTYfns488xQvHv1avvnevSAldt7eLGItKqec/4t\n6QqAiPDZZ3DnneHMvtmoUeHS/d/+BkstBT17hsv4gwbBOeeErvT2zvpFpGtqLf6aAkADAJGOzJ0b\nbl9bsCBsjxwZiv+YMeHWt+22W7jv/Plhzn3zzWHPPcPc+1JLhfl4EYkryZm/BgBoAJAGzemloyt5\nXrAgXKKv5g677RZue4Nw6R7g009hp51gjTVg990XHQCUnT7L8SnHnZP0sr96AEQa3Ntvw0UXwe9/\nH7ar7013D4OC116D/v07XuRGRLIRc86/JV0BECmwf/873N/+ySdhu3dvOOss+M53dKlepGg6W/x1\nBUCkwfzrX+H2ueWWgzffDLfQ9eih++RFiijNM/9mRXgcsORIU1NT1iE0hI7yPHduKP5bbgnTpsF6\n64UGPRX/2umzHJ9yXJssij9oACBSOMOHL5zff+iheE+gE5H4sir+oB4AkUI54wy44AI47rjw32uv\nnXVEItJZ9Sr+6gEQKak5c2D2bPjqV+HJJ+GGG+I9dlZE0pHlmX8zXTyURDSnl457723ivPNg/fXD\nk+z69AnF/+GHVfzrRZ/l+JTj1uWh+IOuAIjkyty58LOfwf/+b1jQ54wz4BvfCM+cF5Hiy0vxB/UA\niGTuk0/glVfgtNPgmWfCa3/6E+y99+LPtReR4opV/NUDIFIg7uERuffdByeeCJ9/Hh6NO2IEDBkS\n7ucXkfLI05l/M/UASCKa0+uaWbPgJz+BZZYJC/f88Iewzz7hqXsvvQT77huKv/Icn3Icn3Ic5LH4\ng64AiER3zTXhkbo9eoRH7kJ4nO6xx4aH8IhIeeW1+IN6AESicg+d/JtsEtbn79ULBg3S4j0ijSCt\n4q/HAaMBgOTH/Plw0EFw++1h+733oG/fbGMSkfSk+lS/Tg4AdB4iiWhOr2Njx0L37qH4P/FEuAqQ\ntPgrz/Epx/E1ao7zfNm/mgYAInV22WWw3XYwfTrssEPW0YhImopS/EFTACJ18dRTcOaZ4T7+6dPh\nxhvh0EOzjkpE0pRV8dcUgEhGRo6E7beHJZcMT+ebMkXFX6TRFOnMv5kGAJJIo87pteeoo2DTTcOi\nPgMHwgordP2YynN8ynF8jZLjIhZ/0DoAIp126aVhCd+33oLXX886GhHJQlGLP6gHQKRmTU2h2E+Z\nArfeCs8+C8cdBwccAHvumXV0IpK2vBR/rQOABgBSf7fdBnfdBXffDZMmwX77hUfz9ugBRx6pLn+R\nRpWX4g9qApSUNMqcHsADD8CBB8Jyy8GwYeH+/n/8Izyp74or4hb/RspzVpTj+Mqa4zwV/65QD4BI\nK0aMCGf7Q4fCxReHDn8RkbIUf9AUgMhivv3t8ACf/fYLZ/wiIpDf4q8pAJEumjcPTjstFP9bb1Xx\nF5GF8lr8u0IDAEmkjHN68+aFW/q6d4ff/AYuuCB09mepjHnOG+U4vrLkuIzFH9QDIMIPfgB//CNc\nfTUccwxY4gtpIlJWZS3+oB4AaXDusMQSofgfe2zW0YhInhSl+KsHQKQTLrwwfD/88GzjEJF8KUrx\n7woNACSRsszpAbz9NpxxBpx7bljYJ0/KlOe8Uo7jK2qOG6H4gwYA0oCmTw/L+q67Lqy2Ghx/fNYR\niUheNErxB/UASIN57DHYZZcw77/bbmG1PxERKG7xVw+ASAcefDAU/332gfnzVfxFZKGiFv+u0ABA\nEinqnN7tt4cn9l1+Odx5Z9bRdKyoeS4S5Ti+ouS4EYs/aAAgDeCKK2D//WGPPcLje3Wfv4g0a9Ti\nD+oBkBKbMQN69w6X+3/xCzjnnKwjEpE8KUvx72wPgFYClNL55BN4551wb//8+TBxIqyxRtZRiUie\nlKX4d4WmACSRPM7pzZ0Ln34KH3wQHuaz1lqw884wdSrce28xi38e81w2ynF8ec2xin+gKwBSWJ9/\nDnvvDQ8/vPC15ZYLa/ufdRb06pVdbCKSTyr+C6kHQArp889hqaXCf//zn+GMv3v3bGMSkXwra/FX\nD4A0lIMOCt/nzMnfMr4ikj9lLf5doR4ASSTrOb0FC+CII8K9/HfdVd7in3WeG4FyHF9ecqzi3zpd\nAZDcmz0bzj4bHn0UXnwRZs0KT/HbZ5+sIxORvFPxb5t6ACTXJkyA9dYLt/P95jew2WZhDf8lNXQV\nkQ40SvFXD4CUzs9/DuefD6uvDk89FW7vExGpRaMU/65QD4Akktac3sSJ8Otfw223wdixjVf88zJ3\nWmbKcXxZ5VjFvza6AiC58tBD8NlnYQGfTTaB/fbLOiIRKRIV/9qpB0By46GHYPfdQ9FfYonQ+LfJ\nJllHJSJF0ajFv7M9ABoASC688QZssEHo7L/rrqyjEZGiadTiD50fAKgHQBKJMac3dSqceCJsvbWK\nfzPNT8enHMeXVo4bufh3hXoAJFPu4cx/8mS46aasoxGRolHx7zxNAUimTj89LOrz8cew0kpZRyMi\nRaLiH2gdACmEjz+G22+H730PevaEmTNDs5+Kv4gkoeLfdeoBkES6Mqe3//6wyirhUb0nnQSTJoUB\nwJln1i++stD8dHzKcXyxcqziXx+6AiBRjRsXGvumTg1n/g89BAMHQq9eWUcmIkWk4l8/6gGQaB5/\nHHbeGXr3hm99CzbdFH7wg6yjEpGiUvFvndYBQAOAvJg1C37xC7j7blh5ZXjsMbDEH00RkYVU/Num\ndQAkFR3N6d1xR2ju++1v4eST4e9/V/HvDM1Px6ccx1evHKv4x6EBgNTNeefB0KFwyikwdy4cdxys\ntlrWUYlIkan4x6MpAOmSN9+E/v1hmWXCpf/rr4fDD886KhEpAxX/2mgdAEndo4/CoEHQpw+MGROa\n/ZZeOuuoRKQMVPzj0xSAJNI8p3fEEaH4/+pXMGECrLqqin89aX46PuU4vs7mWMU/HboCIIn97Gfw\nl7/AI4+EQYCISL2o+KdHPQCSyOOPw0EHQVMTbLxx1tGISJmo+HeOegAkuqeeCgv73HCDir+I1JeK\nf/rUAyA1GTcOBg+Gffdt4pvfzDqa8tP8dHzKcXy15ljFPxsaAEhNbroJ1loLTjgh60hEpExU/LOj\nHgDpkDvsuGN4gt/BB2cdjYiUhYp/fWgpYInmuutgzhw44ICsIxGRslDxz54GANKuyZPhtNPgiiug\nWzfNm6ZFeY5POY6vrRyr+OeDBgDSrlNPhcMOgy23zDoSESkDFf/8UA+AtGrqVBgxAs48E157DZZb\nLuuIRKToVPzjKPQ6AGY2BBhOuCJxtbtf2OL9FYG/AKsD3YCL3f3atONsFO+8A2uvDUsuCffdp+Iv\nIl2n4p8/mU8BmNkSwGXAXsAmwGFmtlGL3X4MjHL3AcBg4GIzy8XgpYxefx1WWAE+/xx2223R9zRv\nmg7lOT7lOL7mHKv451PmAwBgIPCGu09w97nATcDQFvt8CDSfhy4HTHH3eSnG2DDc4Yc/hK23Bkt8\nQUlEZFEq/vmVh7PovsC7VdvvEQYF1a4CHjKz94FewCEpxdZQZs6E734X3n47LPvbGv0fOB3Kc3zK\ncTpU/PMrDwOAWpwBvOTug81sPeBBM9vM3adnHViZHHAA3H8/jB4NK62UdTQiUmQ688+/PAwAJgJr\nVW2vWXmt2g7AeQDu/paZvQ1sBDzf8mBHH300/fr1A6BPnz4MGDDgiw9f83yUthffvuwyeOaZJm6/\nHTbaqO39R40axUknnZR5vGXfrp6fzkM8ZdwePny4/n2ItN3U1MTQoUM54ogjchFP2babmpq49tpr\nAb6od52R+W2AZtYNeB3YDfgAeBY4zN1HV+1zMfCZu//KzFYlFP7N3f2TFsfSbYAJLVgAjzwCu+8e\nzv733LP9/Zuamr74QEo8ynN8ynEcTVVn/qCpljR09jbAzAcA8MVtgL9j4W2AF5jZcYC7+5VmthLw\nJ8KVAgPOd/cbWzmOBgAJvP46HHIIvPRSWOznhhvU+CcinVdd/FX401PoAUC9aABQu8svhx/8APr0\ngYcegi22yDoiESkyFf/s6GFAUrM33gjF/5RT4MMPkxX/5nkoiUt5jk85rp+2ir9ynG95aAKUlF11\nVbjP/3/+J+tIRKTodOZfXJoCaEBmofifckrWkYhIkan454OmAKQml1wCPXrAscdmHYmIFJmKf/Fp\nANBAPv4YfvIT+O1v4b/+q3PH0JxeOpTn+JTjzqu1+CvH+aYBQMn95z/h/v7994dVV4XNN4cf/Ui3\n+4lI5+jMvzzUA1By++4L99wDW24J550H22yjx/uKSOeo+OeT1gFAA4CWZs+GZZaBG2+EQw/NOhoR\nKTIV//xSE6AsYsGC0OnfowcceGD9jqs5vXQoz/Epx7XrbPFXjvNN6wCU0IwZ8PWvhzX+//xn6N49\n64hEpKh05l9emgIooTPPhHPPhQcfDA/5ERHpDBX/YtAUgADw2Wdwyy1w5ZUq/iLSeSr+5acBQImM\nHg29e8OKK8KRR8b5HZrTS4fyHJ9y3LZ6FX/lON/UA1AiJ5wQ7vV/9FHo1i3raESkiHTm3zjUA1AS\nY8fChhvCBx/AaqtlHY2IFJGKfzFpHQAadwAwfTrsvTdMmgSvv551NCJSRCr+xaUmwAa28srwxBNw\nxRXxf5fm9NKhPMenHC8Uq/grx/mmHoCCu+WWsOLftGmhAVBEJAmd+TcuTQEU3K67wmabwfDhWUci\nIkWj4l8O6gGg8QYAzz0X7vWfNAmWWirraESkSFT8y0M9AA1m/HgYOBAOOyzd4q85vXQoz/E1co7T\nKv6NnOMi0ACggF54AdZZB/r00aV/EUlGZ/7STFMABfS1r8Eaa4Suf0t80UdEGpWKfzl1dgpAdwEU\nzJlnwt13w0cfqfiLSO1U/KUlTQEUyMyZ4Sl/V18Nq6ySTQya00uH8hxfI+U4q+LfSDkuIg0ACuSB\nB8Jtf8cem3UkIlIUOvOXtqgHoEB23BG+/nU47bSsIxGRIlDxbwy6DbDkHnkERo6EPfbIOhIRKQIV\nf+mIBgAFceGFMGAAbLFFtnFoTi8dynN8Zc5xXop/mXNcBroLoADmz4f774e//jXrSEQk7/JS/CX/\n1ANQAK++Cv/93zBnjm79E5G2qfg3Jj0LgPIOAA45BObNg1tvzToSEckrFf/GpSbAkpo5E+69Fy6+\nOOtIAs3ppUN5jq9MOc5r8S9TjstIA4Ccu+8+2Hxz6Ncv60hEJI/yWvwl/zQFkHP77gsHHghHHZV1\nJCKSNyr+AuoBAMo3AJg1C3r2hFdegU03zToaEckTFX9pph6AErr44nDpf+ONs45kIc3ppUN5jq/I\nOS5K8S9yjhuB1gHIqfHjw5P/Hn0UunXLOhoRyYuiFH/JP00B5JB7eOBPUxOMHQvdu2cdkYjkgYq/\ntKazUwC6ApBDTzwB114LY8ao+ItIoOIv9aYegJyZMQN23hlOOAE23DDraBanOb10KM/xFSnHRS3+\nRcpxI0p8BcDM1gcOBTYGlnX3/SqvrwlsBjzh7p/VNcoGMno0LL88/O53WUciInlQ1OIv+ZeoB8DM\nTgXOZeHAwd29W+W9/sAY4Mfu/od6B1pjfIXvAbj00nD5/4UXso5ERLKm4i+1iH4boJntD1wAPAns\nCCyyOK27vwG8CAxNGoQstMQSsM02WUchIllT8ZfYkvQAnAyMB4a4+5PA9Fb2eRXI4cx1McyZE64A\n7Lpr1pG0TXN66VCe48tzjstS/POcY0k2ABgA3Ovus9vZ531g1a6F1LiOOgr69oWDD846EhHJSlmK\nv+RfkibAbsDnHeyzUg37SCsuuwxuvhlGjsw6kvbpH6R0KM/x5THHZSv+ZfgbyizJFYC3gG3betPM\nDNgeGN3VoBrNjBnw85/D3/8O22+fdTQikoWyFX/JvyQDgFuAgWb2/TbePwnYCLi5y1E1mMcegy22\ngIMOyjqSjmlOLx3Kc3x5ynFZi3+eciyLSzIFcDFwCPB/ZnYw0B3AzIYBOwGDgFHA7+sbYvk9+yys\nvHLWUYhIFspa/CX/kq4DsBJwObA/0PKew38A33X3T+oXXjJFWwdg3jw47zwYNgxOPx3OPz/riEQk\nTSr+Ug+dXQegUw8DMrO+hH6AFYFPgafdfULiA9VZkQYAw4bBr34V/vuSS8LSv0toYWaRhqHiL/US\nfSGgau4+0d1vdfcr3f3mPBT/IpkwIRT/c86BmTPhpJOKU/w1p5cO5Tm+LHPcKMVfn+N8S7IS4Gdm\ndloH+/zMzD7teljl5R4W+tlwQ/jFL2CZZbKOSETS1CjFX/Kv5ikAM1sADHP3s9vZ5+fA2c3PB0hb\n3qcA5s2D5ZaD2bNh7Fjo3z/riEQkTSr+EkNnpwASPw2wA72BOXU+Zmk89FAo/pMnw4orZh2NiKRJ\nxV/ypt0pADPbovmr8tIa1a9VfW1tZgcChwFvRI+6gMaOhSFD4JBDil38NaeXDuU5vjRz3KjFX5/j\nfOvoCsDzQPM1dQe+W/lqiwFn1SGu0jnoINh5Z7jiiqwjEZE0NWrxl/xrtwfAzC4iFH4DfgI8RXgc\ncEvzgSnAw+7+rwhx1iSvPQDf+x5cdRW8/z6svnrW0YhIWlT8JQ3R1wEws4+BC939oqS/JC15HAC4\nh1v8/ud/4JRTso5GRNKi4i9pib4OgLuvnOfin1ffrUyYnHhitnHUi+b00qE8xxczxyr+gT7H+VaQ\n5WeK6eOP4eqr4dFHoXv3rKMRkTSo+EtRJF4K2Mz2BvYC+gJLtbKLu/vQOsSWWN6mAB5+GM46Cx5/\nPOtIRCQNKv6ShejrAJjZkoQH/nyV0BTY3BzYzKteF2DMGNhoo6yjEJE0qPhL0SSZAjgF2AcYDvQj\nFPtfAxsA3wM+Am4C+tQ3xOL66CPo2zfrKOpLc3rpUJ7jq2eOVfxbp89xviUZABwGvOzuP3X3dyqv\nfe7ub7r7H4HBwFDgqHoHWVRvvQUrrJB1FCISk4q/FFWS2wBnANe4+/GV7fnAr939zKp9bgY2cvfN\nYwRbQ4y56gFYZZXQBPj1r2cdiYjEoOIveZDG44DnA9OrtqcDLRe1fRtYP2kQZTRnTrgLYK+9so5E\nRGJQ8ZeiSzIAmAisWbX9JrBti302BaZ1NagyGDo0PO2vR4+sI6kvzemlQ3mOrys5VvGvjT7H+ZZk\nAPAksE3V9h3A5mb2OzPbxczOAoYADX/T22efwQsvwF/+knUkIlJvKv5SFkl6APYAfgkc4e7jzWw5\nYCThrL/5FsB3gV3cfXyccDuMMRc9AAMHwvjx8N575bsCINLIVPwlj6I/C6CNX7oUcChh3n88cIu7\nf9rpA3ZRHgYAzz8PW28N06fDsstmGoqI1JGKv+RVGk2Ai3H3Oe5+nbuf6e5XZ1n882L4cNh++/IW\nf83ppUN5ji9JjlX8O0ef43yr67MALGjYdQDmzYMbboBjjsk6EhGpFxV/KasuTQEsciCzA4GzCesA\ndKvLQZPHkOkUwNixsOGGMHcuLFnzIssiklcq/lIE0Z4FYGa9gOOArYA5wCPA9e6+oPL+LsDFwH8T\nGgEfSBpEWdx9NwwerOIvUgYq/lJ27U4BmFkf4HngN8AhwJHANcCNlfd/CzwMbAE8Cuzk7kNiBpxn\nl10W5v/LTHN66VCe42svxyr+9aHPcb51dK56GuFhP2MJD/oB+CZwkJndSBgUvAKc7O4PR4uyACZO\nhHHj4Igjso5ERLpCxV8aRbs9AGb2CuHpfhu4+6zKa8sBY4DVgBHAN9x9XgqxdijLHoCDDoLHHw9P\nABSRYlKC3lgeAAAgAElEQVTxlyKKdRvgOsCdzcUfwN3/Q1gFEOD0vBT/rN16K/z2t1lHISKdpeIv\njaajAUBP4MNWXm9+7c36hlNMw4eH7wcckG0cadCcXjqU5/iqc6ziH4c+x/nW1YWAFtQjCDMbYmZj\nzGysmZ3Wxj6DzOxFM/u3mT1Sj99bD1OmwMknw69+Bcssk3U0IpKUir80qo56ABYAf6t8VTsEOKjy\ntdi8g7vfVnMAZksQmgx3A94HngMOdfcxVfv0JjyMaE93n2hmK7n75FaOlXoPwMCB8NZbYSAgIsWi\n4i9lEG0dAODgytdivxO4pY2fSbIQ0EDgDXefAGBmNwFDCY2Gzb4J3OruEwFaK/5ZmDIFnnsO/vWv\nrCMRkaRU/KXRdTQAuI3wpL+Y+hKeItjsPcKgoNoGQPfKpf9ewKXufn3kuDr04IOwxx6wxRZZR5Ke\npqYm/WOZAuU5rqamJoYOHcqIESOU54j0Oc63dgcA7n5QWoF0YEnCYkO7AssCT5nZU+6+WBPi0Ucf\nTb9+/QDo06cPAwYM+OID2NyQUo/tP/0Jjj22icMPB6j/8fO6PWrUqFzFo21tJ90GOPjggzmixaId\neYmvTNv69yLOdlNTE9deey3AF/WuM+r2LIBOB2C2LTCseQVBMzsdcHe/sGqf04Cl3f1Xle0/Ave6\n+60tjpVKD8D8+WG536OPhj/+Ebpl8uQDEUmqSZf9pYQyeRxwnTwHrG9ma5tZD+BQFq4z0GwEsKOZ\ndTOznsA2wOiU4/zCJ5+E7yr+IsWh4i+yqMwHAO4+H/gx4SFCrwI3uftoMzvOzL5X2WcMcD/wMvA0\ncKW7v5ZVzM88A0st1ZjFv/oyqsSjPNdXa8VfOY5POc63XDy3zt3vAzZs8doVLbYvAi5KM662TJkC\n+++fdRQiUgud+Yu0LvMegHpKqwfg178O0wAX5WI4IiJtUfGXRlDkHoDCGTcO1l036yhEpD0q/iLt\n0wCgE8aMgfXWyzqKbGhOLx3Kc9fUUvyV4/iU43zTACCh2bNh5Ejo1SvrSESkNTrzF6lN4h4AMxsM\nfAvYGFjW3QdUXt8A2J2wZO9H9Q60xtii9wC89Rasvz5Mnw7LLhv1V4lIQir+0ohiPgug+pf8HjiO\n8ByAeSy65v9M4H8JjxAubXvcU0/BkCEq/iJ5o+IvkkzNUwBm9h3g+8DNwJrAr6vfd/f3gKeAfeoZ\nYN489hgMHpx1FNnRnF46lOdkOlP8leP4lON8S9IDcBxhoZ7D3f19Wn9I0FigtO1xc+bAX/8KO+2U\ndSQi0kxn/iKdU3MPgJnNIKzAd3Jl+yzgl+7erWqfXwM/cfelYwRbQ4xRewB+8hO45RaYMAEs8WyL\niNSbir9IOj0A84HuHeyzOjAjaRBF4A433QSXXabiL5IHKv4iXZNkCmAMsHNbb5pZd8JzcV/qYky5\n9Oij8MEHWgJYc3rpUJ7bV4/irxzHpxznW5IBwA3ApmZ2XhvvXwCsBfy5y1Hl0GuvwcCBOvsXyZrO\n/EXqI0kPQHfgn8COwDhgFrAJcD2wA6H57wFg71QW5G89xmi/+pJLwtz/8OFRDi8iNVDxF1lc9GcB\nuPtcYC/gEmA1YFPCegBHAmtUXh+aVfGP7eWXwyOARSQbKv4i9ZVoKWB3n+3upwArAFsDQ4DtgBXd\n/afuPidCjLkwciQsv3zWUWRPc3rpUJ4XFaP4K8fxKcf5lmglwGaVqwH/qnMsudarF+y9d9ZRiDQe\nnfmLxJGkB+Bh4E+Etf5nRo2qk2L1AMyfD0suGZ4DoMcAi6RHxV+kY53tAUgyAFhAWP1vBnArcK27\nP5r0F8YUawDw+OOw885hLQARSYeKv0htojcBAv2B84DJwFHAw2b2tpn9ysxKu/wvwMcfw7bbZh1F\nPmhOLx2Nnuc0in+j5zgNynG+JbkL4C13/6W7rwvsSrjff0XgTGCsmT1mZt82s9K1ys2aBeusk3UU\nIo1BZ/4i6ah5CqDVHzbrCRxIuCIwiHBb4Cx371WX6JLHE2UK4Pzz4d134fe/r/uhRaSKir9IcmlM\nASzG3We6+/XA14AzgHnAMl05Zh6NGgU77JB1FCLlpuIvkq4uDQDMbAczuxL4gLAUcHfgmXoElicv\nvABrr511FPmgOb10NFqesyj+jZbjLCjH+ZZ4HQAzW5uw+t+RwLqEy/4TgT8A17n763WNMAfefBP6\n9cs6CpFy0pm/SDaS3AZ4DKHo70S4cjALuB24DngwD0sAx+gBePVV2HRTmDMHevSo66FFGp6Kv0jX\ndbYHIMkVgKsr358kFP2b3f2zpL+waO6/Hw4+WMVfpN5U/EWylaQH4DxgA3ff0d2vaoTiD3DPPbBe\nqVc5SEZzeukoe57zUPzLnuM8UI7zreYrAO5+ZsxA8qpnT9huu6yjECmPPBR/EeniOgB5E6MHYJtt\n4JxzYM8963pYkYak4i9Sf3XvATCzlwlr/+/r7hMq27Vwd988aSB5NG8ePPss9O2bdSQixafiL5Iv\n7fUArFH56la1vXoNX2vECjZt06ZBt26wySZZR5IfmtNLR9nynMfiX7Yc55FynG9tXgFw95Xa224E\nb74JvXtnHYVIseWx+IuIegDa9fjjcMYZ8MQTdTukSENR8ReJL/qzAMzsDjM7tIN9vmFmdyQNIq/m\nzIGllso6CpFiUvEXybck6wB8Ddigg336A/t0Ppx8ee45+PzzrKPIF83ppaPoeS5C8S96jotAOc63\nLj0MqBVLE54IWAqzZsG222YdhUixFKH4i0iyZwEsAM5y93PaeH9F4B5gFXdfp34h1q7ePQCHHQZb\nbgmnnFK3Q4qUmoq/SPqi9ACY2WfNX5WXfl79WtXXDGASsBXw9+Th59NTT8HKK2cdhUgxqPiLFEtH\nUwBjgdcrXw5Mqdqu/noFuA/4GfDzWMGm6e23YcIE2G23rCPJF83ppaNoeS5i8S9ajotIOc63dp8F\n4O5bNf93ZQrgCnc/O3pUOfB//wdbbAFrrpl1JCL5VsTiLyLJegA2ASa5+8dxQ+q8evYADBgAxxwD\nJ55Yl8OJlJKKv0j2OtsDoIWA2jwWjBkDG25Yl8OJlI6Kv0g+xHgY0E8q/3mNu0+r2u6Qu/82aSB5\ncsAB4Xv//tnGkUdNTU36xz4Fec9zGYp/3nNcBspxvrXXA3ARofHvLmBa1XZHowwHCj0AGDECbrwR\nlqj3KgkiJVCG4i8i7UwBmFnzin4Pu/usqu0Oufvd9QguqXpMAcydCz16wMyZsMwydQpMpCRU/EXy\nRz0A1GcAMG0a9OsXvovIQir+IvkU/WFAjeL66+HTT7OOIr90X2868pbnMhb/vOW4jJTjfEvyNMC+\nZrazmfWsem0JM/uZmY00swfMbM84YaZjwgQ44QT46U+zjkQkP8pY/EUk2ToAfwQOBFZ1988rr50B\nnFe12zxgW3d/od6B1qKrUwDbbw9TpsC//w3du9cxMJGCUvEXyb80pgC2Bx6qKv4GnAC8BXwZ2BWY\nA9R8u2DefPghXHedir8IqPiLlF2SAcBqwISq7c2AVYHL3H2MuzcBI4Dt6hdeej76KKz/v9xyWUeS\nb5rTS0fWeW6E4p91jhuBcpxvSQYASwFzq7Z3INzz/1DVaxOA1esQV+oeeCB833jjbOMQyVojFH8R\nSdYD8DrwprvvU9m+kzDfv3LVPlcA+7v7KjGCrSHGTvcAHHEE9OwJV1xR56BECkTFX6R40ugBuA/Y\n08yGmdnpwBDCKoHV+gPvJA0iDx59FHbfPesoRLKj4i/SWJIMAC4APgR+Cfwa+AQY1vymma1ImBZ4\nvI7xpWbppWGzzbKOIv80p5eOtPPciMVfn+X4lON8a+9ZAItw9w/M7MtA85LAD7r7lKpd1gDOJjQC\nFs6cObDUUllHIZK+Riz+IqKlgAFwhxVXhNGjYdVVIwQmklMq/iLFV/fHAXfwy1YANgf6AJ8Co9z9\nk84cKw8mToSpU2GVTFoXRbKh4i/S2BI9C8DMVjOzW4CPgH8CtwAPApPM7BYzK+QtgM88A+usA5Z4\n/NR4NKeXjth5VvHXZzkNynG+1XwFwMxWAkYC6xCaAZ8EPiDc978dcACwhZkNdPfJEWKN5uqrYeut\ns45CJB0q/iICydYBuAQ4ETgXOM/d51S91wP4f4Q7BH7n7idHiLWWGBP3AMydCz16wL33wpAhkQIT\nyQkVf5Hy6WwPQJIBwFvAeHffrZ19HgLWcfd1kwZSD50ZAHzwAfTtC/PmwRJ6OLKUmIq/SDmlsRBQ\nX+DpDvZ5mnA7YGHMnQtrrqniXyvN6aWj3nlW8V+cPsvxKcf5lqTs/QdYs4N9+lb2K4y5c/X0Pyk3\nFX8RaU2SKYARwB7ADu7+YivvbwY8BfzT3YfWNcoadWYKYPRo2H9/GDMmUlAiGVLxFym/NNYBuADY\nG3jKzP4EPEK4C2A1YBBwTOV4FyQNIkuTJkHv3llHIVJ/Kv4i0p6apwDc/SngcOBz4DjgRqAJuAn4\nATAPOLKyX2Gcc44eApSE5vTS0dU8q/h3TJ/l+JTjfEu0EqC7/83MHgQOBrYAehNWAnwR+Ju7T61/\niHFNmgQHHph1FCL1o+IvIrVo+GcB9O8P99wTvosUnYq/SOOJ2gNgZvsDAwEHnnH3Qj7xryV3ePNN\n9QBIOaj4i0gS7fYAmFmPyuI+twCnAqcBt5nZP82s8DfPPVXpVtBDgGqnOb10JM2zin9y+izHpxzn\nW0dNgMcDg4FpwF+AGyr/PRg4IW5o8Y0fDzvtlHUUIl2j4i8indFuD4CZPQtsAHzF3d+tvLY28DLw\nursPTCXKGiXtATjvPHj5Zbj55ohBiUSk4i8isZYC3hC4rbn4A7j7BOC2ynuFNm4c9OqVdRQinaPi\nLyJd0dEAoBfwTiuvv1N5r9CmT4eNNso6imLRnF46Osqzin/X6bMcn3Kcbx0NAAxY0Mrrrb1WKDNn\nwt/+pgGAFI+Kv4jUQ0c9AAuAKytf1Y4DvgNsRRgkLMLdX6hjjDVL0gNw661w0EEwYwb07Bk5MJE6\nUfEXkZY62wNQywCgrR2sjffc3ROtMFgvSQYAJ54IkyfDDTdEDkqkTlT8RaQ1sZoAX2jn619tvL7Y\nkwLz6NJL4StfyTqK4tGcXjpa5lnFv/70WY5POc63ds/U3X2rtAJJ04JKB8OPfpRtHCK1UPEXkRhy\n8SwAMxsCDCdckbja3S9sY7+tgSeBQ9z9tlber2kK4Jpr4NvfDgMBS3zRRCQ9Kv4i0pFYUwDRmdkS\nwGXAXsAmwGFmtlhvfmW/C4D7u/o7x42Dww5T8Zd8U/EXkZgyHwAQHjL0hrtPcPe5wE3A0Fb2O57w\nTIJJXf2FY8bAl7/c1aM0Js3ppWP48OEq/pHpsxyfcpxveRgA9AXerdp+r/LaF8xsDWA/d/8Drdx2\nmNRzz8Hmm3f1KCJxNDU1cdZZZ6n4i0hUmdyu1wnDCU8ibNbmIODoo4+mX79+APTp04cBAwZ88Y9o\nU1MT48bBO+8MYqutFo5Oq9/XdsfbzfIST5m2R40axXnnnceIESO+eC1P8ZVpu/m1vMRT1u1meYmn\nDNtNTU1ce+21AF/Uu87IvAnQzLYFhrn7kMr26YS1BC6s2mdc838CKwEzgO+5+x0tjtVhE+B998H5\n58Ojj9bxjxCpgybN+YtIJxS2CRB4DljfzNY2sx7AocAihd3d1618rUPoA/hhy+Jfq08/heWX73LM\nDavlqF7qo2XxV57jU47jU47zLfMpAHefb2Y/Bh5g4W2Ao83suPC2t1yGuEuXLK66ClZaqStHEKkv\nnfmLSBYSTwGY2fqEs/SNgWXdfb/K62sCmwFPuPtn9Q60xtg6nALo1g0efhh22SWloETaoeIvIl0V\n5VkArfySU4FzWXjlwN29W+W9/sAY4MeVbv3UdTQAeOUV2GwzLQAk+aDiLyL1EL0HwMz2JyzE8ySw\nI3Bx9fvu/gbhOQCt3cOfC48/DjvuqOLfFZrTq4+Oir/yHJ9yHJ9ynG9JmgBPBsYDQ9z9SWB6K/u8\nCmxYh7iimDUL+vfPOgppdDrzF5E8SDIAGADc6+6z29nnfWDVroUUz+zZsNpqWUdRbCpYXVNr8Vee\n41OO41OO8y3JAKAb8HkH+6xUwz6ZefZZWGqprKOQRqUzfxHJkyQDgLeAbdt608wM2B4Y3dWgYpg5\nE+64A3baKetIik1zep2TtPgrz/Epx/Epx/mWZABwCzDQzL7fxvsnARsBN3c5qgjmzg0LAO26a9aR\nSKPRmb+I5FHNtwGa2bLAs4Qi3wR0B3YAzgF2AgYBo4Dt3D2TaYD2bgP88EPYeGOYOjXloKShqfiL\nSGydvQ2w5pUA3X2Gme0CXA7sz8IH8vyy8v0fwHezKv4dGTEC9tor6yikkaj4i0ieJXoWgLtPdveD\ngLWAg4HvA4cB67j7ge7+SYQY6+KFF2D77bOOovg0p1ebrhZ/5Tk+5Tg+5TjfOvUsAHefCNxa51ii\n+ve/4VvfyjoKaQQ68xeRIsj8ccD11F4PwBprwDPPwJe+lHJQ0lBU/EUkbdGfBWBml9Z4THf3E5MG\nUg9tDQDmzAl3AMycGR4GJBKDir+IZCGNAcCCDnZxQmPgFw8ISltbA4ApU2DDDWHy5AyCKpmmpiYV\nt1bUu/grz/Epx/Epx+mIfhcA8JU2Xu8DbA2cDjxCeFpgrsyZAz16ZB2FlJXO/EWkiOrWA2Bm6wIv\nAd939xvqctDkMbR6BWD8eBg0KHwXqScVfxHJWvTHAXfE3ccBI4Cf1uuY9TJzJiyzTNZRSNmo+ItI\nkdVtAFDxAWGlwFx5/31YffWsoygH3dcbxC7+ynN8ynF8ynG+1W0AUHkY0M7Af+p1zHoZNw7WXjvr\nKKQsdOYvImWQ5C6ALdp4a0ngS8C3gb2A69z92PqEl0xbPQCHHQabbw6nn55BUFIqKv4ikjdp3AXw\nPOFWvzZjqOzzs6RBxNbUBN/4RtZRSNGp+ItImSS5AnARrQ8AFgBTCU8KfKTNpfhS0NoVgHfeCZf/\np0yBFVbIKLASadT7etMu/o2a5zQpx/Epx+lI42mApyQ9eB6MHAmbbqriL52nM38RKaOkSwGPdvc/\nxA2p81q7AvDXv8Kdd8KNN2YUlBSair+I5F0a6wAcBxSul37uXOjePesopIhU/EWkzJIMAN4BVowV\nSCwaANRXo9zXm3Xxb5Q8Z0k5jk85zrckA4Cbgb3MbLlYwcQwYYKeACjJZF38RUTSkKQHYGngLqAX\n4cE/z7n7jIixJdZaD8Cpp8KCBXDRRRkFJYWi4i8iRZPGOgCTCFcMegIPVX7pTBa/NdDdvXfSQGIZ\nOxZ22y3rKKQIVPxFpJEkmQIYC4wG/gW8UPkaA7ze4mtsnWPsklmz4L/+K+soyqOsc3p5K/5lzXOe\nKMfxKcf5lmQdgK1iBhLL/Pmw6qpZRyF5lrfiLyKShnZ7AMzsSGCUu7+cXkid11oPwGabwWWXwc47\nZxSU5JqKv4gUXax1AK4F9utURDkwbx688gqstlrWkUgeqfiLSCOr2+OA82jaNDCDDTbIOpLyKMuc\nXt6Lf1nynGfKcXzKcb6VegDw7ruw/PJZRyF5k/fiLyKSho56ABYAw9z97PRC6ryWPQB/+xucey68\nXIgOBkmDir+IlE3MdQD6mNlaSQ7q7u8kDSSGOXNCE6AIqPiLiFSrZQrgRODtBF/jokTaCbNnQ48e\nWUdRLkWd0yta8S9qnotEOY5POc63Wq4AfAZMix1IDGPGhIcBSWMrWvEXEUlDqXsAzjgjNAGecUaG\nQUmmVPxFpOxirQNQaHPmaAqgkan4i4i0rdQDgAkTNACot6LM6RW9+Bclz0WmHMenHOdbqQcATz8N\nX/5y1lFI2ope/EVE0lDaHgB3WGYZ+OADPQ2wkaj4i0ijibIOgLsX9grBc89Bz57Qp0/WkUhaVPxF\nRGpX2ALfkcmToV+/8CwAqZ+8zumVrfjnNc9lohzHpxznW2kHAP/5T7gCIOVXtuIvIpKGdnsAiqa6\nB+Dii+Gll+DPf844KIlKxV9EGp3WAWjhmmugV6+so5CYVPxFRDqvtAOAzz6DQw/NOoryycucXtmL\nf17yXGbKcXzKcb6VdgDw3nvQv3/WUUgMZS/+IiJpKG0PQLduYSngJWt53JEUhoq/iMii1ANQZd48\nWLBAxb9sVPxFROqnlAOA997LOoLyympOr9GKv+ZO41OO41OO862UA4AxY2DDDbOOQuql0Yq/iEga\nStkDcOGFcNdd8PjjWUckXaXiLyLSPvUAVPn0U9hoo6yjkK5S8RcRiaeUA4D334eNN846inJKa06v\n0Yu/5k7jU47jU47zrZQDgEmTwqOApZgavfiLiKShlD0Ae+wBp54Ke+yRdUSSlIq/iEgy6gGo8uab\nsPLKWUchSan4i4ikp3QDgLlzwzoAm22WdSTlFGtOT8V/UZo7jU85jk85zrfSDQCmTIEVVoAlSveX\nlZeKv4hI+krXA/Dyy86hh8Krr2YdjdRCxV9EpGvUA1AxebLm/4tCxV9EJDulGwCMHw9rrpl1FOVV\nrzk9Ff/2ae40PuU4PuU430o3AHjnHVhnnayjkPao+IuIZK90PQBnnul06wZnnZV1NNIaFX8RkfpS\nD0DF/PnQrVvWUUhrVPxFRPJDAwBJpLNzeir+yWjuND7lOD7lON80AJDoVPxFRPKndD0AJ5/s9O0L\nP/1p1tEIqPiLiMSmHoAKXQHIDxV/EZH8Kt0AYMwYLQMcU61zeir+XaO50/iU4/iU43wrXal86CFY\nZZWso2hsKv4iIvlXuh4AcN5/H1ZfPetoGpOKv4hIutQDUGXFFbOOoDGp+IuIFEfpBgBHHQU9emQd\nRXm1Naen4l9fmjuNTzmOTznOt9INALbcMusIGo+Kv4hI8ZSuB+DSS53jj886ksah4i8ikq1C9wCY\n2RAzG2NmY83stFbe/6aZvVT5esLMvtLWsWbNihurLKTiLyJSXJkPAMxsCeAyYC9gE+AwM9uoxW7j\ngJ3dfXPgXOCqto633nqxIhVYOKen4h+X5k7jU47jU47zLfMBADAQeMPdJ7j7XOAmYGj1Du7+tLt/\nWtl8Gujb1sEs8UUQSUrFX0Sk+PIwAOgLvFu1/R7tFHjgO8C9bb2pVQDjU/GPT7mNTzmOTznOtyWz\nDiAJMxsMHAPs2PY+6cXTaHTmLyJSHnkYAEwE1qraXrPy2iLMbDPgSmCIu09t62CXXno0L77YD4A+\nffowYMCAL4pV83yUtpNvNzU1MXToUI444ohcxFP27eq50zzEU8bt4cOH69+HyNujRo3ipJNOyk08\nZdluamri2muvBaBfv350Vua3AZpZN+B1YDfgA+BZ4DB3H121z1rAQ8AR7v50O8fyO+90vva1yEE3\nmKaqM3/QZb00NDU1Kc+RKcfxKcfp6OxtgJkPACDcBgj8jtCTcLW7X2BmxwHu7lea2VXAAcAEwIC5\n7j6wleP43Xc7X/1qmtGXW3Xx1/+RRUTyp9ADgHoxM7/nHmfvvbOOpBxU/EVE8q/QCwHV0xKl+4uy\n0Vbxb56HkriU5/iU4/iU43wrXbnUXQBdpzN/EZHyK90UwIMPOrvvnnUkxaXiLyJSLJoCqNAVgM5T\n8RcRaRylGwCoB6Bzai3+mtNLh/Icn3Icn3Kcb6Url7oCkJzO/EVEGk/pegCampxddsk6kuJQ8RcR\nKTb1AFR86UtZR1AcKv4iIo2rdAOAddbJOoJi6Gzx15xeOpTn+JTj+JTjfCvdAEA9AB3Tmb+IiJSu\nB6BMf08MKv4iIuWiHgDpkIq/iIg00wCgQdSr+GtOLx3Kc3zKcXzKcb5pANAAdOYvIiItqQeg5FT8\nRUTKTT0AshgVfxERaYsGACUVq/hrTi8dynN8ynF8ynG+aQBQQjrzFxGRjqgHoGRU/EVEGot6AETF\nX0REaqYBQEmkVfw1p5cO5Tk+5Tg+5TjfNAAoAZ35i4hIUuoBKDgVfxGRxqYegAak4i8iIp2lAUBB\nZVX8NaeXDuU5PuU4PuU43zQAKCCd+YuISFepB6BgVPxFRKSaegAagIq/iIjUiwYABZGX4q85vXQo\nz/Epx/Epx/mmAUAB5KX4i4hIeagHIOdU/EVEpD3qASghFX8REYlFA4Ccymvx15xeOpTn+JTj+JTj\nfNMAIIfyWvxFRKQ81AOQMyr+IiKShHoASkDFX0RE0qIBQE4UpfhrTi8dynN8ynF8ynG+aQCQA0Up\n/iIiUh7qAciYir+IiHSFegAKSMVfRESyogFARopa/DWnlw7lOT7lOD7lON80AMhAUYu/iIiUh3oA\nUqbiLyIi9aQegAJQ8RcRkbzQACAlZSn+mtNLh/Icn3Icn3KcbxoApKAsxV9ERMpDPQCRqfiLiEhM\n6gHIIRV/ERHJKw0AIilr8decXjqU5/iU4/iU43zTACCCshZ/EREpD/UA1JmKv4iIpEk9ADmg4i8i\nIkWhAUCdNErx15xeOpTn+JTj+JTjfNMAoA4apfiLiEh5qAegi1T8RUQkS+oByICKv4iIFJUGAJ3U\nqMVfc3rpUJ7jU47jU47zTQOATmjU4i8iIuWhHoCEVPxFRCRP1AOQAhV/EREpCw0AaqTiH2hOLx3K\nc3zKcXzKcb5pAFADFX8RESkb9QB0QMVfRETyTD0AEaj4i4hIWWkA0AYV/9ZpTi8dynN8ynF8ynG+\naQDQChV/EREpO/UAtKDiLyIiRaIegDpQ8RcRkUahAUCFin9tNKeXDuU5PuU4PuU43zQAQMVfREQa\nT8P3AKj4i4hIkakHoBNU/EVEpFE17ABAxb9zNKeXDuU5PuU4PuU43xpyAKDiLyIija7hegBU/EVE\npEzUA1ADFX8REZGgYQYAKv71oTm9dCjP8SnH8SnH+dYQAwAVfxERkUWVvgdAxV9ERMpMPQCtUPEX\nERFpXWkHACr+cWhOLx3Kc3zKcXzKcb7lYgBgZkPMbIyZjTWz09rY51Ize8PMRpnZgPaOp+Ifz6hR\no1BTjaoAAA0HSURBVLIOoSEoz/Epx/Epx/mW+QDAzJYALgP2AjYBDjOzjVrsszewnrv3B44DLm/r\neCr+cU2bNi3rEBqC8hyfchyfcpxvmQ8AgIHAG+4+wd3nAjcBQ1vsMxT4M4C7PwP0NrNVWzuYir+I\niEjH8jAA6Au8W7X9XuW19vaZ2Mo+ACr+kY0fPz7rEBqC8hyfchyfcpxvmd8GaGYHAnu5+/cq24cD\nA939hKp97gTOd/cnK9v/BE519xdaHKs89zSKiIjUqDO3AS4ZI5CEJgJrVW2vWXmt5T5f6mCfTiVA\nRESkEeVhCuA5YH0zW9vMegCHAne02OcO4EgAM9sWmObuH6UbpoiISHlkfgXA3eeb2Y+BBwgDkqvd\nfbSZHRfe9ivd/R4z+6qZvQnMAI7JMmYREZGiy7wHQERERNKXhymAxOq9cJAsrqMcm9k3zeylytcT\nZvaVLOIsslo+x5X9tjazuWZ2QJrxlUWN/14MMrMXzezfZvZI2jEWXQ3/XqxoZvdW/j1+xcyOziDM\nQjOzq83sIzN7uZ19ktU9dy/UF2HQ8iawNtAdGAVs1GKfvYG7K/+9DfB01nEX6avGHG8L9K789xDl\nuP45rtrvIeAu4ICs4y7aV42f5d7Aq0DfyvZKWcddpK8ac3wW4U4ugJWAKcCSWcdepC9gR2AA8HIb\n7yeue0W8AlDXhYOkVR3m2N2fdvdPK5tP08a6DNKmWj7HAMcDtwCT0gyuRGrJ8zeBW919IoC7T045\nxqKrJccfAstV/ns5YIq7z0sxxsJz9yeAqe3skrjuFXEAUNeFg6RVteS42neAe6NGVD4d5tjM1gD2\nc/c/ALrFtXNq+SxvAKxgZo+Y2XNmdkRq0ZVDLTm+CtjEzN4HXgJOTCm2RpK47mV+F4AUm5kNJtyV\nsWPWsZTQcKB6PlWDgDiWBLYAdgWWBZ4ys6fc/c1swyqVM4CX3H2wma0HPGhmm7n79KwDa2RFHADU\nbeEgaVMtOcbMNgOuBIa4e3uXpmRxteR4K+AmMzPCvOneZjbX3VuukyFtqyXP7wGT3X02MNvMHgM2\nJ8xrS8dqyfEOwHkA7v6Wmb0NbAQ8n0qEjSFx3SviFIAWDoqvwxyb2VrArcAR7v5WBjEWXYc5dvd1\nK1/rEPoAfqjin1gt/16MAHY0s25m1pPQQDU65TiLrJYcjwZ2B6jMS28AjEs1ynIw2r4SmLjuFe4K\ngGvhoOhqyTFwJrAC8PvKGepcdx+YXdTFUmOOF/mR1IMsgRr/vRhjZvcDLwPzgSvd/bUMwy6UGj/L\n5wN/MrOXCAXsVHf/JLuoi8fM/goMAlY0s3cId1b0oAt1TwsBiYiINKAiTgGIiIhIF2kAICIi0oA0\nABAREWlAGgCIiIg0IA0AREREGpAGACIiIg1IAwCRhMxsNzNbYGb/L+tY8qKyiM4CM3sg4c/9pfJz\na8SKTURapwGAlFKlqLT3dWTWMdaDmZ3Tyt82o/Js9v9NubA6LRYsMrNzKzFt387PLIgeWY3M7Nut\n5HO2mY03sz+b2aZ1+j0aRErmCrcSoEgCDgyj9aUzR6UbSlQOPAI8VtleGdgL+BFwsJkNdPd3ogYQ\nVoPbmLACWcvY2ltt7BTgHMLjYvPkBRYuZ9ub8LCrw4EDzWzXyuNWRQpNAwApNXc/J+sYUvKwu/+6\necPMugEPArsAPweOix2Au49t5eV2n2BYWas8j8/peMHdz65+wcyuAr4NnAvs0cXj68mOkjlNAUjD\nM7MNzOzCyrPgJ1Uu+b5tZpcnuYRuZuua2R/N7E0zm2lmk83sZTP7vZn1bmX/b1WeQT/VzGaZ2atm\ndoaZde/q3+Tu8wnPYDdgkWc0mNnqZvaHymXt2Wb2kZndYmYDWomxh5mdZGYvmNknZja9kpt/WHgU\ndPN+i/UAmNm7QPMl7ieqLql/XrXPIj0AZrZDZfvmtv42MxtbmeZYvsXre5vZvZW8z67873ChmS2X\nKHltu7ryfatWYqr5M2Rm1xPWzXfg3Kq8zG85VRLzMyKiKwAicDDwHcJl9CeAucBXgO8C+5jZVh09\nVavyj/zzQE/gHuDvwDLAOsARwHDg06r9r6u8PqGy76fA9oRHpg42s7286w/qaD7L/OI4ZrYuMBJY\nBfgncAPhUa4HV/7W/dz9/qpj/AU4CHgJuBaYDfQFdiKcBT/Szu+/GNivsu81QPM0xPyqfRaZInD3\nkWb2FvB1M+vt7p9W7YuZbQesD9zo7p9VvX428AtgMuHS/ceER/r+DNjLzLZ395ntxJrEnFZeS/IZ\nupWQgyOBh1k4dQMLc5TWZ0QambvrS1+l+yI0ls0nPDGr5df/b+/MQ7UqogD+Oy5lqaCFueBGJRRS\npFGJZlqSFllESVhhzzJsgRaqP9I0DQlRSXAFbbMEjbJAgkRLEy0qXB4lbWpopkmiJUYu9fD0xzkf\nXq/3+u7ney+he34w3PedOXdm7nzzvjkzc2ZuTUq3C9AyI41hnsaslHyIpz8+IXvadR/NSOd84JzE\n54f9/qVJuce95Ok8VvA5p7j++JS8BbDW4+Yn5Ktd9lxKfwBQh03Ht3JZe9f9PCfv9om/m/szrcop\nX/+cNBZ7fJeEbILLxmboL/C4oQnZzZ73WqBNSv8hj5tWsD7HuP7CjLg3K99bRlyD21AqvtHaSIQI\neeGsFyBChKYICQMgK6ypIp1vgR9SsjwD4DgwukCaW4Aj6c7K45oDvwOfFSzfFM93NScMnDnANpfv\nBbq7bg+XbQeaZaS1xOtnpH9u5/qfFihHYxoA3bMMD+Bcr5vd+JtMXf6h6/fKyeMbYHfB+qwYABsT\n9TkTe+f9cey99l2rbIuF2lBTtZEIEfJCLAEE/2tUtXkRPd8W+ABwJTbyTd6X9mzPYjnW2S0QkduA\nlVgH9n0qnzZAb8zr/VmRU3zBBJtmv7xIuRMM9gDwNzaVPBeYqqp7Xd7Hr+tUNWvr3RpgpOu9o6oH\nRWQFcIuI1GJT1+uBr1T1aJXlK4yq7hKRtcBgEblUVbd71J2YUbJAVZNT3/2wafn7cuqzBdBZRNqq\n6p8Fi9GHE/VVYRswSFX3Zd3QCG2okk5TtZEgOIkwAILSIyJzsC1ze4AVfq10cGOAzvWloao7RORa\nbMQ4DLjLkpZdwAxVneeqF/i1I/DiaZL8p4pHUGCCqk6tR6/iiLg3J74ib5eQ3Q08D9yLTT0LcERE\n3sOWEfZXUc5qWATcCNQAE11Wgz3r2yndSp2erj4VaAMUNQBeV9WxACJyEbaWPwX4SEQGqOpJfgCN\n0YYSNEUbCYJTCAMgKDUi0gl4HKgFBqRHtlLFgUE+2h8pIs0wB7ShwBPAbBE5pKqLOeEIuEFV+zXG\nM1SKWkCnknennPjOKT28PiYDk0WkK3AD8CA20u2GTWU3Be8D8zAnuIki0hFb69+YnlUBDgHHVDXv\nuRqEj/hfFpELsaWeycC4SnxjtiGnqdpIEJxEbAMMys4lWOe5KuOHuwfQs9oEVfW4qtaq6jTs8BjB\npq9R82r/EbgivY3tP6DWrwMlY14ZuAkbKW/OullVd6vqEsyw2YFN0de3xa7i8V9oKSaR12FgGdDN\ntxuO8jTeylD/EuggIr2qyeMMmAQcAJ5yY6jCmbSh3Ho5y20kKBFhAARlZ6dfB/rIHQDv2BZS8H9E\nRK7O6Qwro9LkGvBMbIvgG1k/8CLSPmtPfkNR1Z+xbWqXAE+m8hwA3INto1vusg4i0jsjqbZAa2wK\nuq6ebA9gnWP3MyjyIr+3BjMAjmFe8Wlmut5rPho/CRFp7cszDcL9B2YArTBjoMJOv1bThg74Na9e\nzkobCcpFLAEEpUZV94jIMmyte7OIfIKtlQ/F1ou3UMzZajQwRkTWAz8BB7H96rdj3tyzE3m+KiJ9\ngbHAID88Zxe29nsxtm9+IalO+jRUc6rcI5gj30wRuRXYhO0OGIF16KNV9Yjrdgc2iMjXWD3sxupm\nOHbc8CsJ3TzWYLMK073DOggcL+CvgKquE5EdmP9BS+ADVf0jQ+9jEXkBW6Pf5o6LO7A1/57YaYhr\ngDvqy7MAc4FngBoRmaaq28+wDX2H+VzcLyKKff8KLFLVPU3QRoLgVM72NoQIEZoiYFus6grqnocd\nrrIVOIyN6GZhP+LrsfXlpP4QbAp3XEJ2HTAfe8fAfmzEvxU7je+ynHyHY1vYfsNGt78CX2BrzJlb\n2jLSyDwHoJ57unhZd2KOavuwg2b6pvTaYQ54q4FfMENmj38ekdJt7uVYmZHfKGz54S/XOZaIW4zN\nInTJKeskv6cOGF7Pc10PvIsZKke9XjcB04GrCtbNGM9vwWl0Kmc+LE3IqmpDfs81XpcHObFFtX9K\np8FtJEKEvCCqcZBUEARBEJSN8AEIgiAIghISBkAQBEEQlJAwAIIgCIKghIQBEARBEAQlJAyAIAiC\nICghYQAEQRAEQQkJAyAIgiAISkgYAEEQBEFQQsIACIIgCIIS8i/NNqsWE5AahQAAAABJRU5ErkJg\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x134994050>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr_rf, tpr_rf, thresholds = roc_curve(Y_test, rf_preds[:,1])\n",
    "\n",
    "figsize(8,8)\n",
    "plt.plot(np.arange(0.,1.,1e-5),np.arange(0.,1.,1e-5), c='black')\n",
    "plt.plot(fpr_rf,tpr_rf, c='b')\n",
    "plt.title('ROC curve', size=20)\n",
    "plt.xlabel('False Positive Rate', size=20)\n",
    "plt.ylabel('True Positive Rate', size=20)\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Качество Gradient Boosting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC 0.931830401361\n",
      "accuracy 0.874405036082\n"
     ]
    }
   ],
   "source": [
    "print 'AUC', roc_auc_score(Y_test, bst_preds[:,1])\n",
    "print 'accuracy', accuracy_score(Y_test, bst.predict(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgAAAAIGCAYAAAAiITfnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nXgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1346d1950>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr_gb, tpr_gb, thresholds = roc_curve(Y_test, bst_preds[:,1])\n",
    "\n",
    "figsize(8,8)\n",
    "plt.plot(np.arange(0.,1.,1e-5),np.arange(0.,1.,1e-5), c='black')\n",
    "plt.plot(fpr_gb,tpr_gb, c='b')\n",
    "plt.title('ROC curve', size=20)\n",
    "plt.xlabel('False Positive Rate', size=20)\n",
    "plt.ylabel('True Positive Rate', size=20)\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Кросс-валидация"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.cross_validation import ShuffleSplit, KFold\n",
    "from tqdm import tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "cv = KFold(X.shape[0], n_folds=5, shuffle=True, random_state=123)\n",
    "\n",
    "lr_aucs = []\n",
    "lr_accs = []\n",
    "rf_aucs = []\n",
    "rf_accs = []\n",
    "bst_aucs = []\n",
    "bst_accs = []\n",
    "for train_ids, test_ids in tqdm(cv):\n",
    "    X_train, Y_train = X[train_ids, :], Y[train_ids, :]\n",
    "    X_test, Y_test = X[test_ids, :], Y[test_ids, :]\n",
    "    lr.fit(X_train, Y_train.flatten())\n",
    "    lr_aucs.append(roc_auc_score(Y_test, lr.predict_proba(X_test)[:,1]))\n",
    "    lr_accs.append(accuracy_score(Y_test, lr.predict(X_test)))\n",
    "    \n",
    "    rf.fit(X_train, Y_train.flatten())\n",
    "    rf_aucs.append(roc_auc_score(Y_test, rf.predict_proba(X_test)[:,1]))\n",
    "    rf_accs.append(accuracy_score(Y_test, rf.predict(X_test)))\n",
    "    \n",
    "    bst.fit(X_train, Y_train.flatten())\n",
    "    bst_aucs.append(roc_auc_score(Y_test, bst.predict_proba(X_test)[:,1]))\n",
    "    bst_accs.append(accuracy_score(Y_test, bst.predict(X_test)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Смотрим как себя ведёт качество на разных фолдах"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "figsize(16,8)\n",
    "\n",
    "plt.subplot(121)\n",
    "plt.ylabel('AUC', size=20)\n",
    "plt.xlabel('fold number', size=20)\n",
    "plt.plot(range(5), lr_aucs, label='Logistic Regression')\n",
    "plt.plot(range(5), rf_aucs, label='Random Forest')\n",
    "plt.plot(range(5), bst_aucs, label='Gradient Boosting')\n",
    "plt.legend()\n",
    "\n",
    "plt.subplot(122)\n",
    "plt.ylabel('accuracy', size=20)\n",
    "plt.xlabel('fold number', size=20)\n",
    "plt.plot(range(5), lr_accs, label='Logistic Regression')\n",
    "plt.plot(range(5), rf_accs, label='Random Forest')\n",
    "plt.plot(range(5), bst_accs, label='Gradient Boosting')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "ids = numpy.arange(X.shape[0])\n",
    "numpy.random.shuffle(ids)\n",
    "\n",
    "train_len = int(len(ids) * 0.8)\n",
    "dtrain = xgb.DMatrix(X[ids[:train_len],:], label=Y[ids[:train_len],:])\n",
    "dtest = xgb.DMatrix(X[ids[train_len:],:], label=Y[ids[train_len:],:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[0]\teval-auc:0.889302\ttrain-auc:0.895943\n",
      "[1]\teval-auc:0.895084\ttrain-auc:0.900388\n",
      "[2]\teval-auc:0.898256\ttrain-auc:0.904709\n",
      "[3]\teval-auc:0.898086\ttrain-auc:0.904308\n",
      "[4]\teval-auc:0.900235\ttrain-auc:0.907075\n",
      "[5]\teval-auc:0.900605\ttrain-auc:0.907007\n",
      "[6]\teval-auc:0.901868\ttrain-auc:0.908684\n",
      "[7]\teval-auc:0.901973\ttrain-auc:0.908967\n",
      "[8]\teval-auc:0.902364\ttrain-auc:0.909490\n",
      "[9]\teval-auc:0.903633\ttrain-auc:0.910443\n",
      "[10]\teval-auc:0.904618\ttrain-auc:0.911046\n",
      "[11]\teval-auc:0.905298\ttrain-auc:0.911705\n",
      "[12]\teval-auc:0.906209\ttrain-auc:0.912846\n",
      "[13]\teval-auc:0.906309\ttrain-auc:0.913592\n",
      "[14]\teval-auc:0.907000\ttrain-auc:0.914280\n",
      "[15]\teval-auc:0.907157\ttrain-auc:0.914550\n",
      "[16]\teval-auc:0.907308\ttrain-auc:0.914721\n",
      "[17]\teval-auc:0.907828\ttrain-auc:0.915327\n",
      "[18]\teval-auc:0.907826\ttrain-auc:0.915462\n",
      "[19]\teval-auc:0.908082\ttrain-auc:0.915814\n",
      "[20]\teval-auc:0.908105\ttrain-auc:0.916054\n",
      "[21]\teval-auc:0.908421\ttrain-auc:0.916444\n",
      "[22]\teval-auc:0.908342\ttrain-auc:0.916513\n",
      "[23]\teval-auc:0.908493\ttrain-auc:0.916710\n",
      "[24]\teval-auc:0.908601\ttrain-auc:0.916841\n",
      "[25]\teval-auc:0.908829\ttrain-auc:0.916975\n",
      "[26]\teval-auc:0.909039\ttrain-auc:0.917327\n",
      "[27]\teval-auc:0.909262\ttrain-auc:0.917832\n",
      "[28]\teval-auc:0.909605\ttrain-auc:0.918073\n",
      "[29]\teval-auc:0.909789\ttrain-auc:0.918442\n",
      "[30]\teval-auc:0.909810\ttrain-auc:0.918704\n",
      "[31]\teval-auc:0.910238\ttrain-auc:0.919341\n",
      "[32]\teval-auc:0.910424\ttrain-auc:0.919553\n",
      "[33]\teval-auc:0.910352\ttrain-auc:0.919797\n",
      "[34]\teval-auc:0.910376\ttrain-auc:0.920019\n",
      "[35]\teval-auc:0.910867\ttrain-auc:0.920650\n",
      "[36]\teval-auc:0.911115\ttrain-auc:0.920989\n",
      "[37]\teval-auc:0.911288\ttrain-auc:0.921193\n",
      "[38]\teval-auc:0.911426\ttrain-auc:0.921456\n",
      "[39]\teval-auc:0.911537\ttrain-auc:0.921706\n",
      "[40]\teval-auc:0.911668\ttrain-auc:0.921969\n",
      "[41]\teval-auc:0.911866\ttrain-auc:0.922112\n",
      "[42]\teval-auc:0.911979\ttrain-auc:0.922434\n",
      "[43]\teval-auc:0.912460\ttrain-auc:0.922940\n",
      "[44]\teval-auc:0.912704\ttrain-auc:0.923199\n",
      "[45]\teval-auc:0.912872\ttrain-auc:0.923456\n",
      "[46]\teval-auc:0.913012\ttrain-auc:0.923691\n",
      "[47]\teval-auc:0.913224\ttrain-auc:0.924034\n",
      "[48]\teval-auc:0.913548\ttrain-auc:0.924467\n",
      "[49]\teval-auc:0.913865\ttrain-auc:0.924790\n",
      "[50]\teval-auc:0.914078\ttrain-auc:0.925071\n",
      "[51]\teval-auc:0.914316\ttrain-auc:0.925291\n",
      "[52]\teval-auc:0.914486\ttrain-auc:0.925467\n",
      "[53]\teval-auc:0.914806\ttrain-auc:0.925858\n",
      "[54]\teval-auc:0.915003\ttrain-auc:0.926075\n",
      "[55]\teval-auc:0.915201\ttrain-auc:0.926341\n",
      "[56]\teval-auc:0.915513\ttrain-auc:0.926747\n",
      "[57]\teval-auc:0.915740\ttrain-auc:0.926991\n",
      "[58]\teval-auc:0.915867\ttrain-auc:0.927127\n",
      "[59]\teval-auc:0.916160\ttrain-auc:0.927435\n",
      "[60]\teval-auc:0.916260\ttrain-auc:0.927639\n",
      "[61]\teval-auc:0.916485\ttrain-auc:0.927978\n",
      "[62]\teval-auc:0.916673\ttrain-auc:0.928156\n",
      "[63]\teval-auc:0.916806\ttrain-auc:0.928378\n",
      "[64]\teval-auc:0.917041\ttrain-auc:0.928658\n",
      "[65]\teval-auc:0.917254\ttrain-auc:0.928910\n",
      "[66]\teval-auc:0.917407\ttrain-auc:0.929150\n",
      "[67]\teval-auc:0.917498\ttrain-auc:0.929320\n",
      "[68]\teval-auc:0.917724\ttrain-auc:0.929517\n",
      "[69]\teval-auc:0.917782\ttrain-auc:0.929732\n",
      "[70]\teval-auc:0.917988\ttrain-auc:0.929940\n",
      "[71]\teval-auc:0.918200\ttrain-auc:0.930229\n",
      "[72]\teval-auc:0.918310\ttrain-auc:0.930398\n",
      "[73]\teval-auc:0.918417\ttrain-auc:0.930582\n",
      "[74]\teval-auc:0.918615\ttrain-auc:0.930759\n",
      "[75]\teval-auc:0.918816\ttrain-auc:0.930969\n",
      "[76]\teval-auc:0.918939\ttrain-auc:0.931060\n",
      "[77]\teval-auc:0.918897\ttrain-auc:0.931161\n",
      "[78]\teval-auc:0.919014\ttrain-auc:0.931313\n",
      "[79]\teval-auc:0.919079\ttrain-auc:0.931581\n",
      "[80]\teval-auc:0.919155\ttrain-auc:0.931786\n",
      "[81]\teval-auc:0.919311\ttrain-auc:0.931945\n",
      "[82]\teval-auc:0.919458\ttrain-auc:0.932089\n",
      "[83]\teval-auc:0.919557\ttrain-auc:0.932250\n",
      "[84]\teval-auc:0.919639\ttrain-auc:0.932444\n",
      "[85]\teval-auc:0.919796\ttrain-auc:0.932702\n",
      "[86]\teval-auc:0.919983\ttrain-auc:0.932858\n",
      "[87]\teval-auc:0.920115\ttrain-auc:0.932962\n",
      "[88]\teval-auc:0.920132\ttrain-auc:0.933097\n",
      "[89]\teval-auc:0.920255\ttrain-auc:0.933305\n",
      "[90]\teval-auc:0.920392\ttrain-auc:0.933424\n",
      "[91]\teval-auc:0.920563\ttrain-auc:0.933645\n",
      "[92]\teval-auc:0.920673\ttrain-auc:0.933753\n",
      "[93]\teval-auc:0.920752\ttrain-auc:0.933933\n",
      "[94]\teval-auc:0.920856\ttrain-auc:0.934026\n",
      "[95]\teval-auc:0.920930\ttrain-auc:0.934287\n",
      "[96]\teval-auc:0.921108\ttrain-auc:0.934555\n",
      "[97]\teval-auc:0.921269\ttrain-auc:0.934664\n",
      "[98]\teval-auc:0.921334\ttrain-auc:0.934919\n",
      "[99]\teval-auc:0.921429\ttrain-auc:0.935006\n",
      "[100]\teval-auc:0.921568\ttrain-auc:0.935170\n",
      "[101]\teval-auc:0.921625\ttrain-auc:0.935270\n",
      "[102]\teval-auc:0.921727\ttrain-auc:0.935329\n",
      "[103]\teval-auc:0.921869\ttrain-auc:0.935593\n",
      "[104]\teval-auc:0.921885\ttrain-auc:0.935757\n",
      "[105]\teval-auc:0.921970\ttrain-auc:0.935823\n",
      "[106]\teval-auc:0.922087\ttrain-auc:0.935902\n",
      "[107]\teval-auc:0.922181\ttrain-auc:0.936157\n",
      "[108]\teval-auc:0.922241\ttrain-auc:0.936234\n",
      "[109]\teval-auc:0.922343\ttrain-auc:0.936380\n",
      "[110]\teval-auc:0.922383\ttrain-auc:0.936551\n",
      "[111]\teval-auc:0.922469\ttrain-auc:0.936626\n",
      "[112]\teval-auc:0.922591\ttrain-auc:0.936720\n",
      "[113]\teval-auc:0.922568\ttrain-auc:0.936855\n",
      "[114]\teval-auc:0.922664\ttrain-auc:0.936920\n",
      "[115]\teval-auc:0.922763\ttrain-auc:0.937068\n",
      "[116]\teval-auc:0.922817\ttrain-auc:0.937101\n",
      "[117]\teval-auc:0.922876\ttrain-auc:0.937304\n",
      "[118]\teval-auc:0.922988\ttrain-auc:0.937427\n",
      "[119]\teval-auc:0.923041\ttrain-auc:0.937580\n",
      "[120]\teval-auc:0.923117\ttrain-auc:0.937655\n",
      "[121]\teval-auc:0.923213\ttrain-auc:0.937744\n",
      "[122]\teval-auc:0.923320\ttrain-auc:0.937830\n",
      "[123]\teval-auc:0.923359\ttrain-auc:0.937886\n",
      "[124]\teval-auc:0.923427\ttrain-auc:0.937968\n",
      "[125]\teval-auc:0.923492\ttrain-auc:0.938042\n",
      "[126]\teval-auc:0.923602\ttrain-auc:0.938132\n",
      "[127]\teval-auc:0.923655\ttrain-auc:0.938179\n",
      "[128]\teval-auc:0.923665\ttrain-auc:0.938288\n",
      "[129]\teval-auc:0.923754\ttrain-auc:0.938366\n",
      "[130]\teval-auc:0.923781\ttrain-auc:0.938598\n",
      "[131]\teval-auc:0.923865\ttrain-auc:0.938660\n",
      "[132]\teval-auc:0.923980\ttrain-auc:0.938733\n",
      "[133]\teval-auc:0.924034\ttrain-auc:0.938770\n",
      "[134]\teval-auc:0.924151\ttrain-auc:0.938865\n",
      "[135]\teval-auc:0.924241\ttrain-auc:0.938957\n",
      "[136]\teval-auc:0.924290\ttrain-auc:0.938980\n",
      "[137]\teval-auc:0.924388\ttrain-auc:0.939063\n",
      "[138]\teval-auc:0.924429\ttrain-auc:0.939270\n",
      "[139]\teval-auc:0.924502\ttrain-auc:0.939333\n",
      "[140]\teval-auc:0.924518\ttrain-auc:0.939371\n",
      "[141]\teval-auc:0.924567\ttrain-auc:0.939482\n",
      "[142]\teval-auc:0.924623\ttrain-auc:0.939533\n",
      "[143]\teval-auc:0.924665\ttrain-auc:0.939563\n",
      "[144]\teval-auc:0.924747\ttrain-auc:0.939642\n",
      "[145]\teval-auc:0.924817\ttrain-auc:0.939703\n",
      "[146]\teval-auc:0.924837\ttrain-auc:0.939804\n",
      "[147]\teval-auc:0.924904\ttrain-auc:0.939886\n",
      "[148]\teval-auc:0.924972\ttrain-auc:0.939931\n",
      "[149]\teval-auc:0.925066\ttrain-auc:0.940116\n",
      "[150]\teval-auc:0.925110\ttrain-auc:0.940145\n",
      "[151]\teval-auc:0.925147\ttrain-auc:0.940186\n",
      "[152]\teval-auc:0.925166\ttrain-auc:0.940213\n",
      "[153]\teval-auc:0.925251\ttrain-auc:0.940266\n",
      "[154]\teval-auc:0.925235\ttrain-auc:0.940352\n",
      "[155]\teval-auc:0.925314\ttrain-auc:0.940408\n",
      "[156]\teval-auc:0.925330\ttrain-auc:0.940435\n",
      "[157]\teval-auc:0.925362\ttrain-auc:0.940474\n",
      "[158]\teval-auc:0.925413\ttrain-auc:0.940635\n",
      "[159]\teval-auc:0.925488\ttrain-auc:0.940718\n",
      "[160]\teval-auc:0.925521\ttrain-auc:0.940738\n",
      "[161]\teval-auc:0.925523\ttrain-auc:0.940776\n",
      "[162]\teval-auc:0.925545\ttrain-auc:0.940793\n",
      "[163]\teval-auc:0.925629\ttrain-auc:0.940850\n",
      "[164]\teval-auc:0.925668\ttrain-auc:0.941008\n",
      "[165]\teval-auc:0.925673\ttrain-auc:0.941141\n",
      "[166]\teval-auc:0.925724\ttrain-auc:0.941173\n",
      "[167]\teval-auc:0.925741\ttrain-auc:0.941208\n",
      "[168]\teval-auc:0.925788\ttrain-auc:0.941236\n",
      "[169]\teval-auc:0.925850\ttrain-auc:0.941323\n",
      "[170]\teval-auc:0.925891\ttrain-auc:0.941378\n",
      "[171]\teval-auc:0.925934\ttrain-auc:0.941399\n",
      "[172]\teval-auc:0.925929\ttrain-auc:0.941527\n",
      "[173]\teval-auc:0.925988\ttrain-auc:0.941604\n",
      "[174]\teval-auc:0.926051\ttrain-auc:0.941658\n",
      "[175]\teval-auc:0.926081\ttrain-auc:0.941676\n",
      "[176]\teval-auc:0.926174\ttrain-auc:0.941731\n",
      "[177]\teval-auc:0.926219\ttrain-auc:0.941766\n",
      "[178]\teval-auc:0.926189\ttrain-auc:0.941810\n",
      "[179]\teval-auc:0.926149\ttrain-auc:0.941965\n",
      "[180]\teval-auc:0.926176\ttrain-auc:0.942025\n",
      "[181]\teval-auc:0.926215\ttrain-auc:0.942056\n",
      "[182]\teval-auc:0.926255\ttrain-auc:0.942099\n",
      "[183]\teval-auc:0.926304\ttrain-auc:0.942128\n",
      "[184]\teval-auc:0.926318\ttrain-auc:0.942145\n",
      "[185]\teval-auc:0.926360\ttrain-auc:0.942176\n",
      "[186]\teval-auc:0.926372\ttrain-auc:0.942221\n",
      "[187]\teval-auc:0.926432\ttrain-auc:0.942319\n",
      "[188]\teval-auc:0.926456\ttrain-auc:0.942392\n",
      "[189]\teval-auc:0.926495\ttrain-auc:0.942413\n",
      "[190]\teval-auc:0.926468\ttrain-auc:0.942431\n",
      "[191]\teval-auc:0.926466\ttrain-auc:0.942568\n",
      "[192]\teval-auc:0.926489\ttrain-auc:0.942583\n",
      "[193]\teval-auc:0.926534\ttrain-auc:0.942613\n",
      "[194]\teval-auc:0.926556\ttrain-auc:0.942627\n",
      "[195]\teval-auc:0.926567\ttrain-auc:0.942654\n",
      "[196]\teval-auc:0.926603\ttrain-auc:0.942676\n",
      "[197]\teval-auc:0.926623\ttrain-auc:0.942741\n",
      "[198]\teval-auc:0.926614\ttrain-auc:0.942771\n",
      "[199]\teval-auc:0.926669\ttrain-auc:0.942876\n",
      "[200]\teval-auc:0.926618\ttrain-auc:0.943033\n",
      "[201]\teval-auc:0.926659\ttrain-auc:0.943061\n",
      "[202]\teval-auc:0.926711\ttrain-auc:0.943088\n",
      "[203]\teval-auc:0.926735\ttrain-auc:0.943105\n",
      "[204]\teval-auc:0.926761\ttrain-auc:0.943172\n",
      "[205]\teval-auc:0.926787\ttrain-auc:0.943185\n",
      "[206]\teval-auc:0.926818\ttrain-auc:0.943208\n",
      "[207]\teval-auc:0.926856\ttrain-auc:0.943373\n",
      "[208]\teval-auc:0.926852\ttrain-auc:0.943496\n",
      "[209]\teval-auc:0.926864\ttrain-auc:0.943541\n",
      "[210]\teval-auc:0.926884\ttrain-auc:0.943554\n",
      "[211]\teval-auc:0.926912\ttrain-auc:0.943578\n",
      "[212]\teval-auc:0.926945\ttrain-auc:0.943604\n",
      "[213]\teval-auc:0.926961\ttrain-auc:0.943618\n",
      "[214]\teval-auc:0.926938\ttrain-auc:0.943717\n",
      "[215]\teval-auc:0.927017\ttrain-auc:0.943885\n",
      "[216]\teval-auc:0.927051\ttrain-auc:0.943936\n",
      "[217]\teval-auc:0.927072\ttrain-auc:0.944000\n",
      "[218]\teval-auc:0.927103\ttrain-auc:0.944021\n",
      "[219]\teval-auc:0.927117\ttrain-auc:0.944030\n",
      "[220]\teval-auc:0.927102\ttrain-auc:0.944102\n",
      "[221]\teval-auc:0.927129\ttrain-auc:0.944114\n",
      "[222]\teval-auc:0.927117\ttrain-auc:0.944220\n",
      "[223]\teval-auc:0.927147\ttrain-auc:0.944236\n",
      "[224]\teval-auc:0.927215\ttrain-auc:0.944395\n",
      "[225]\teval-auc:0.927266\ttrain-auc:0.944414\n",
      "[226]\teval-auc:0.927258\ttrain-auc:0.944458\n",
      "[227]\teval-auc:0.927278\ttrain-auc:0.944527\n",
      "[228]\teval-auc:0.927265\ttrain-auc:0.944656\n",
      "[229]\teval-auc:0.927245\ttrain-auc:0.944750\n",
      "[230]\teval-auc:0.927270\ttrain-auc:0.944770\n",
      "[231]\teval-auc:0.927311\ttrain-auc:0.944791\n",
      "[232]\teval-auc:0.927317\ttrain-auc:0.944813\n",
      "[233]\teval-auc:0.927343\ttrain-auc:0.944882\n",
      "[234]\teval-auc:0.927376\ttrain-auc:0.944903\n",
      "[235]\teval-auc:0.927397\ttrain-auc:0.945002\n",
      "[236]\teval-auc:0.927396\ttrain-auc:0.945028\n",
      "[237]\teval-auc:0.927396\ttrain-auc:0.945045\n",
      "[238]\teval-auc:0.927409\ttrain-auc:0.945053\n",
      "[239]\teval-auc:0.927353\ttrain-auc:0.945154\n",
      "[240]\teval-auc:0.927362\ttrain-auc:0.945243\n",
      "[241]\teval-auc:0.927408\ttrain-auc:0.945261\n",
      "[242]\teval-auc:0.927416\ttrain-auc:0.945362\n",
      "[243]\teval-auc:0.927424\ttrain-auc:0.945390\n",
      "[244]\teval-auc:0.927361\ttrain-auc:0.945493\n",
      "[245]\teval-auc:0.927402\ttrain-auc:0.945508\n",
      "[246]\teval-auc:0.927452\ttrain-auc:0.945657\n",
      "[247]\teval-auc:0.927477\ttrain-auc:0.945674\n",
      "[248]\teval-auc:0.927474\ttrain-auc:0.945699\n",
      "[249]\teval-auc:0.927439\ttrain-auc:0.945754\n",
      "[250]\teval-auc:0.927467\ttrain-auc:0.945768\n",
      "[251]\teval-auc:0.927469\ttrain-auc:0.945843\n",
      "[252]\teval-auc:0.927473\ttrain-auc:0.945862\n",
      "[253]\teval-auc:0.927484\ttrain-auc:0.945870\n",
      "[254]\teval-auc:0.927520\ttrain-auc:0.945891\n",
      "[255]\teval-auc:0.927529\ttrain-auc:0.945923\n",
      "[256]\teval-auc:0.927534\ttrain-auc:0.945930\n",
      "[257]\teval-auc:0.927542\ttrain-auc:0.945998\n",
      "[258]\teval-auc:0.927557\ttrain-auc:0.946010\n",
      "[259]\teval-auc:0.927603\ttrain-auc:0.946090\n",
      "[260]\teval-auc:0.927624\ttrain-auc:0.946227\n",
      "[261]\teval-auc:0.927618\ttrain-auc:0.946273\n",
      "[262]\teval-auc:0.927632\ttrain-auc:0.946280\n",
      "[263]\teval-auc:0.927653\ttrain-auc:0.946290\n",
      "[264]\teval-auc:0.927642\ttrain-auc:0.946334\n",
      "[265]\teval-auc:0.927663\ttrain-auc:0.946405\n",
      "[266]\teval-auc:0.927669\ttrain-auc:0.946426\n",
      "[267]\teval-auc:0.927681\ttrain-auc:0.946474\n",
      "[268]\teval-auc:0.927690\ttrain-auc:0.946481\n",
      "[269]\teval-auc:0.927684\ttrain-auc:0.946550\n",
      "[270]\teval-auc:0.927674\ttrain-auc:0.946567\n",
      "[271]\teval-auc:0.927656\ttrain-auc:0.946615\n",
      "[272]\teval-auc:0.927676\ttrain-auc:0.946633\n",
      "[273]\teval-auc:0.927683\ttrain-auc:0.946712\n",
      "[274]\teval-auc:0.927704\ttrain-auc:0.946722\n",
      "[275]\teval-auc:0.927718\ttrain-auc:0.946754\n",
      "[276]\teval-auc:0.927723\ttrain-auc:0.946764\n",
      "[277]\teval-auc:0.927739\ttrain-auc:0.946834\n",
      "[278]\teval-auc:0.927759\ttrain-auc:0.946842\n",
      "[279]\teval-auc:0.927770\ttrain-auc:0.946940\n",
      "[280]\teval-auc:0.927739\ttrain-auc:0.947054\n",
      "[281]\teval-auc:0.927739\ttrain-auc:0.947063\n",
      "[282]\teval-auc:0.927752\ttrain-auc:0.947078\n",
      "[283]\teval-auc:0.927764\ttrain-auc:0.947128\n",
      "[284]\teval-auc:0.927793\ttrain-auc:0.947149\n",
      "[285]\teval-auc:0.927831\ttrain-auc:0.947217\n",
      "[286]\teval-auc:0.927846\ttrain-auc:0.947223\n",
      "[287]\teval-auc:0.927830\ttrain-auc:0.947298\n",
      "[288]\teval-auc:0.927835\ttrain-auc:0.947361\n",
      "[289]\teval-auc:0.927844\ttrain-auc:0.947377\n",
      "[290]\teval-auc:0.927866\ttrain-auc:0.947394\n",
      "[291]\teval-auc:0.927878\ttrain-auc:0.947406\n",
      "[292]\teval-auc:0.927881\ttrain-auc:0.947447\n",
      "[293]\teval-auc:0.927893\ttrain-auc:0.947453\n",
      "[294]\teval-auc:0.927879\ttrain-auc:0.947469\n",
      "[295]\teval-auc:0.927889\ttrain-auc:0.947483\n",
      "[296]\teval-auc:0.927911\ttrain-auc:0.947561\n",
      "[297]\teval-auc:0.927933\ttrain-auc:0.947616\n",
      "[298]\teval-auc:0.927932\ttrain-auc:0.947625\n",
      "[299]\teval-auc:0.927943\ttrain-auc:0.947634\n"
     ]
    }
   ],
   "source": [
    "param = {'bst:max_depth':6, 'bst:eta':0.05, 'objective':'binary:logistic' }\n",
    "param['eval_metric'] = 'auc'\n",
    "\n",
    "evallist  = [(dtest,'eval'), (dtrain,'train')]\n",
    "num_round = 300\n",
    "bst = xgb.train(param, dtrain, num_round, evallist)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Качество XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC 0.927942508053\n",
      "accuracy 0.869491785659\n"
     ]
    }
   ],
   "source": [
    "labels = dtest.get_label()\n",
    "preds = bst.predict(dtest)\n",
    "\n",
    "print 'AUC', roc_auc_score(labels, preds)\n",
    "print 'accuracy', accuracy_score(labels, np.array(preds > 0.5, dtype=int))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgAAAAIGCAYAAAAiITfnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIYQQooRIARBCCCFKiBQAIYQQooT8B+pVolFTtvCdAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x13430a5d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fpr_xgb, tpr_xgb, thresholds = roc_curve(labels, preds)\n",
    "\n",
    "figsize(8,8)\n",
    "plt.plot(np.arange(0.,1.,1e-5),np.arange(0.,1.,1e-5), c='black')\n",
    "plt.plot(fpr_xgb,tpr_xgb, c='b')\n",
    "plt.title('ROC curve', size=20)\n",
    "plt.xlabel('False Positive Rate', size=20)\n",
    "plt.ylabel('True Positive Rate', size=20)\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### RoC кривые разных моделей на одном графике"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgAAAAIGCAYAAAAiITfnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nHPI/bo/VfOAl/L0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x133fb5810>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize(8,8)\n",
    "plt.plot(np.arange(0.,1.,1e-5),np.arange(0.,1.,1e-5), c='black')\n",
    "plt.plot(fpr_lr,tpr_lr, label = 'Logistic Regression')\n",
    "plt.plot(fpr_rf,tpr_rf, label = 'Random Forest')\n",
    "plt.plot(fpr_gb,tpr_gb, label = 'Gradient Boosting')\n",
    "plt.plot(fpr_xgb,tpr_xgb, label = 'XGBoost')\n",
    "plt.title('ROC curve', size=20)\n",
    "plt.xlabel('False Positive Rate', size=20)\n",
    "plt.ylabel('True Positive Rate', size=20)\n",
    "plt.legend()\n",
    "plt.grid()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature importance"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Посмотрим на важность признаков в модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x133d86590>"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAngAAAGXCAYAAAAgbS3wAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nERERES3zAsgYsBIw8I5UAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1342467d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize(10,6)\n",
    "plt.plot([i[1] for i in sorted(bst.get_fscore().iteritems(), key = lambda s : -s[1])])\n",
    "plt.title('Feature importances plot', size=20)\n",
    "plt.xlabel('sorted feature number', size=20)\n",
    "plt.ylabel('importance', size=20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Кросс валидация для XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "dtrain_cv = xgb.DMatrix(X, label = Y)\n",
    "\n",
    "param = {'bst:max_depth':6, 'bst:eta':0.05, 'objective':'binary:logistic' }\n",
    "param['eval_metric'] = 'auc'\n",
    "\n",
    "num_round = 300\n",
    "cv_results = xgb.cv(param, dtrain_cv, num_round, as_pandas=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Выведем результаты кросс-валидации на последних 10 итерациях"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "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>test-auc-mean</th>\n",
       "      <th>test-auc-std</th>\n",
       "      <th>train-auc-mean</th>\n",
       "      <th>train-auc-std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>290</th>\n",
       "      <td>0.928893</td>\n",
       "      <td>0.002024</td>\n",
       "      <td>0.948755</td>\n",
       "      <td>0.000750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>291</th>\n",
       "      <td>0.928900</td>\n",
       "      <td>0.002033</td>\n",
       "      <td>0.948790</td>\n",
       "      <td>0.000783</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>292</th>\n",
       "      <td>0.928922</td>\n",
       "      <td>0.002032</td>\n",
       "      <td>0.948853</td>\n",
       "      <td>0.000755</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>293</th>\n",
       "      <td>0.928932</td>\n",
       "      <td>0.002020</td>\n",
       "      <td>0.948920</td>\n",
       "      <td>0.000723</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>294</th>\n",
       "      <td>0.928942</td>\n",
       "      <td>0.002024</td>\n",
       "      <td>0.948969</td>\n",
       "      <td>0.000697</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>295</th>\n",
       "      <td>0.928943</td>\n",
       "      <td>0.002022</td>\n",
       "      <td>0.949021</td>\n",
       "      <td>0.000691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>296</th>\n",
       "      <td>0.928949</td>\n",
       "      <td>0.002018</td>\n",
       "      <td>0.949061</td>\n",
       "      <td>0.000712</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>297</th>\n",
       "      <td>0.928955</td>\n",
       "      <td>0.002022</td>\n",
       "      <td>0.949103</td>\n",
       "      <td>0.000742</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298</th>\n",
       "      <td>0.928953</td>\n",
       "      <td>0.002003</td>\n",
       "      <td>0.949167</td>\n",
       "      <td>0.000673</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>299</th>\n",
       "      <td>0.928956</td>\n",
       "      <td>0.001994</td>\n",
       "      <td>0.949215</td>\n",
       "      <td>0.000648</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     test-auc-mean  test-auc-std  train-auc-mean  train-auc-std\n",
       "290       0.928893      0.002024        0.948755       0.000750\n",
       "291       0.928900      0.002033        0.948790       0.000783\n",
       "292       0.928922      0.002032        0.948853       0.000755\n",
       "293       0.928932      0.002020        0.948920       0.000723\n",
       "294       0.928942      0.002024        0.948969       0.000697\n",
       "295       0.928943      0.002022        0.949021       0.000691\n",
       "296       0.928949      0.002018        0.949061       0.000712\n",
       "297       0.928955      0.002022        0.949103       0.000742\n",
       "298       0.928953      0.002003        0.949167       0.000673\n",
       "299       0.928956      0.001994        0.949215       0.000648"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_results.iloc[290:,:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Кривая обучения"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAngAAAGJCAYAAAAZsU4bAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n8z+3e7FXEvDf4AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x133a3a8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(cv_results['train-auc-mean'], label = 'train')\n",
    "plt.plot(cv_results['test-auc-mean'], label = 'test')\n",
    "plt.xlabel('n_iter', size=20)\n",
    "plt.ylabel('auc', size=20)\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Визуализация точек на карте: pygeoplot\n",
    "\n",
    "[пример с камерами видеонаблюдения](http://nbviewer.jupyter.org/github/romovpa/pygeoplot/blob/master/ipynb/DemoMoscowCCTV.ipynb)"
   ]
  },
  {
   "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": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}