hw1-kmeans-features-Lozinskii.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Алгоритмы интеллектуальной обработки больших объемов данных\n",
    "## Домашнее задание №1 - Метод k-средних, предобработка признаков\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# <hr\\>\n",
    "**Общая информация**\n",
    "\n",
    "**Срок сдачи:** до 7 октября 2017, 06:00   \n",
    "**Штраф за опоздание:** -2 балла после 06:00 7 октября, -4 балла после 06:00 14 октября, -6 баллов после 06:00 21 октября\n",
    "\n",
    "При отправлении ДЗ указывайте фамилию в названии файла   \n",
    "\n",
    "\n",
    "Присылать ДЗ необходимо в виде ссылки на свой github репозиторий в slack @alkhamush\n",
    "Необходимо в slack создать таск в приватный чат:   \n",
    "/todo Фамилия Имя *ссылка на гитхаб* @alkhamush   \n",
    "Пример:   \n",
    "/todo Ксения Стройкова https://github.com/stroykova/spheremailru/stroykova_hw1.ipynb @alkhamush   \n",
    "\n",
    "Используйте данный Ipython Notebook при оформлении домашнего задания.\n",
    "<hr\\>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Первое задание - указать фамилию в имени файла"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "plt.style.use('ggplot')\n",
    "plt.rcParams['figure.figsize'] = (12,5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Пролог: Модуль sklearn\n",
    "На наших практических занятиях мы будем использовать [sklearn](http://scikit-learn.org/stable/index.html) (scikit-learn) - наиболее популярный модуль для работы с методами машинного обучения и анализа данных.\n",
    "\n",
    "Этот модуль регулярно обновляется, расширяется и сопровождается примерами использования тех или иных методов. Помимо самих алгоритмов (например k-средних или линейной регрессии) в нем есть множество вспомогательных функций по работе с данными."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Туториал: Работа с различными типами признаков"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Вспомним какие типы признаков могут встретится при описании объектов исследования и как их представить в доступной для обработки компьтером форме"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Типы признаков мы будем рассматривать на примере довольно популярного набора данных [titanic](https://www.kaggle.com/c/titanic/download/train.csv) - информации о выживаемости пассажиров после известной трагедии корабля Titanic."
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "ARIABLE DESCRIPTIONS:\n",
    "survival        Survival\n",
    "                (0 = No; 1 = Yes)\n",
    "pclass          Passenger Class\n",
    "                (1 = 1st; 2 = 2nd; 3 = 3rd)\n",
    "name            Name\n",
    "sex             Sex\n",
    "age             Age\n",
    "sibsp           Number of Siblings/Spouses Aboard\n",
    "parch           Number of Parents/Children Aboard\n",
    "ticket          Ticket Number\n",
    "fare            Passenger Fare\n",
    "cabin           Cabin\n",
    "embarked        Port of Embarkation\n",
    "                (C = Cherbourg; Q = Queenstown; S = Southampton)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>PassengerId</th>\n",
       "      <th>Survived</th>\n",
       "      <th>Pclass</th>\n",
       "      <th>Name</th>\n",
       "      <th>Sex</th>\n",
       "      <th>Age</th>\n",
       "      <th>SibSp</th>\n",
       "      <th>Parch</th>\n",
       "      <th>Ticket</th>\n",
       "      <th>Fare</th>\n",
       "      <th>Cabin</th>\n",
       "      <th>Embarked</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>Cumings, Mrs. John Bradley (Florence Briggs Th...</td>\n",
       "      <td>female</td>\n",
       "      <td>38.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>PC 17599</td>\n",
       "      <td>71.2833</td>\n",
       "      <td>C85</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>Futrelle, Mrs. Jacques Heath (Lily May Peel)</td>\n",
       "      <td>female</td>\n",
       "      <td>35.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>113803</td>\n",
       "      <td>53.1000</td>\n",
       "      <td>C123</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>McCarthy, Mr. Timothy J</td>\n",
       "      <td>male</td>\n",
       "      <td>54.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>17463</td>\n",
       "      <td>51.8625</td>\n",
       "      <td>E46</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>11</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>Sandstrom, Miss. Marguerite Rut</td>\n",
       "      <td>female</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>PP 9549</td>\n",
       "      <td>16.7000</td>\n",
       "      <td>G6</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>Bonnell, Miss. Elizabeth</td>\n",
       "      <td>female</td>\n",
       "      <td>58.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>113783</td>\n",
       "      <td>26.5500</td>\n",
       "      <td>C103</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    PassengerId  Survived  Pclass  \\\n",
       "1             2         1       1   \n",
       "3             4         1       1   \n",
       "6             7         0       1   \n",
       "10           11         1       3   \n",
       "11           12         1       1   \n",
       "\n",
       "                                                 Name     Sex   Age  SibSp  \\\n",
       "1   Cumings, Mrs. John Bradley (Florence Briggs Th...  female  38.0      1   \n",
       "3        Futrelle, Mrs. Jacques Heath (Lily May Peel)  female  35.0      1   \n",
       "6                             McCarthy, Mr. Timothy J    male  54.0      0   \n",
       "10                    Sandstrom, Miss. Marguerite Rut  female   4.0      1   \n",
       "11                           Bonnell, Miss. Elizabeth  female  58.0      0   \n",
       "\n",
       "    Parch    Ticket     Fare Cabin Embarked  \n",
       "1       0  PC 17599  71.2833   C85        C  \n",
       "3       0    113803  53.1000  C123        S  \n",
       "6       0     17463  51.8625   E46        S  \n",
       "10      1   PP 9549  16.7000    G6        S  \n",
       "11      0    113783  26.5500  C103        S  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_titanic = pd.read_csv('train.csv', sep=',').dropna()\n",
    "df_titanic.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Бинарные признаки\n",
    "Это признаки с двумя возможными значениями\n",
    "* Пол (мужской, женский)\n",
    "* Выжил (да, нет)\n",
    "\n",
    "На вход в алгоритмы обычно подаются в виде значений `{0, 1}` (реже в виде `{-1, 1}`)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    123\n",
       "0     60\n",
       "Name: Survived, dtype: int64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Здесь все нормально\n",
    "df_titanic.Survived.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "male      95\n",
       "female    88\n",
       "Name: Sex, dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Тут стоит выполнить переобознаение\n",
    "df_titanic.Sex.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    95\n",
       "0    88\n",
       "Name: Sex, dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Например так\n",
    "df_titanic.Sex = df_titanic.Sex.replace({'male': 1, 'female': 0})\n",
    "df_titanic.Sex.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Вещественные признаки\n",
    "\n",
    "* Возраст\n",
    "* Сумма, потраченная на билет\n",
    "* Количество родственников на борту\n",
    "\n",
    "Как правило остаются в том же виде или выполняется **нормализация** признаков, т.е. переход к одинаковой шкале, например:\n",
    "\n",
    "* $\\tilde{x} = \\frac{x - \\text{mean}(x)}{\\text{std}(x)}$\n",
    "* $\\tilde{x} = \\frac{x - \\text{min}(x)}{\\text{max}(x) - \\text{min}(x)}$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Категориальные признаки\n",
    "\n",
    "Принимают значения из *неупорядоченного* множества\n",
    "\n",
    "* Порт погрузки (C = Cherbourg; Q = Queenstown; S = Southampton)\n",
    "* Точка отправления (St Louis, MO; Montreal, PQ; .. }\n",
    "* Точка назначения (Chesterville, ON; Roseville, IL; .. }\n",
    "\n",
    "Обычно такие признаки кодируют с помощью так называемого One Hot Encoding: </br>\n",
    "для каждой категории создается бинарная маска-столбец, в которой стоит `1` если значение признака равно значению соответствующей категории категории.\n",
    "\n",
    "Например, для признака \"порт погрзки\" создается бинарная макска из трех стобцов для каждого из портов: С, Q, S<br/>\n",
    "Если пассажир заходил на кораблю в порту Southampton, то его признаковое описание будет соответственно: `[0, 0, 1]`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1     C\n",
       "3     S\n",
       "6     S\n",
       "10    S\n",
       "11    S\n",
       "Name: Embarked, dtype: object"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_titanic.Embarked.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>Embarked=C</th>\n",
       "      <th>Embarked=Q</th>\n",
       "      <th>Embarked=S</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Embarked=C  Embarked=Q  Embarked=S\n",
       "1            1           0           0\n",
       "3            0           0           1\n",
       "6            0           0           1\n",
       "10           0           0           1\n",
       "11           0           0           1"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Сделать в python это можно несколькими способами:\n",
    "\n",
    "# 1: Через pandas (не очень рекомендуется)\n",
    "emb = pd.get_dummies(df_titanic.Embarked, prefix='Embarked', prefix_sep='=')\n",
    "emb.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 1.,  0.,  0.],\n",
       "       [ 0.,  0.,  1.],\n",
       "       [ 0.,  0.,  1.],\n",
       "       [ 0.,  0.,  1.],\n",
       "       [ 0.,  0.,  1.]])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 2: Функционал sklearn\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "\n",
    "encoder = OneHotEncoder(sparse=False)\n",
    "\n",
    "emb = df_titanic.loc[:, ['Embarked']].replace({'C': 0, 'Q': 1, 'S': 2}).values\n",
    "emb_enc = encoder.fit_transform(emb)\n",
    "emb_enc[:5]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Про другием способы работы с категориальными переменными можно прочитать вот [тут](https://alexanderdyakonov.wordpress.com/2016/08/03/python-%D0%BA%D0%B0%D1%82%D0%B5%D0%B3%D0%BE%D1%80%D0%B8%D0%B0%D0%BB%D1%8C%D0%BD%D1%8B%D0%B5-%D0%BF%D1%80%D0%B8%D0%B7%D0%BD%D0%B0%D0%BA%D0%B8/)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Порядковый признаки"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Принимают значения из *упорядоченного* множества\n",
    "\n",
    "* Класс каюты (1 > 2 > 3)\n",
    "\n",
    "В зависимости от метода остаются можно оставить в исходном виде или применить OneHotEncoding."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Текстовые признаки - Bag of words"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Для демонстрации возьмем текст стихотворения Блока:\n",
    "\n",
    "Ночь, улица, фонарь, аптека,<br/>\n",
    "Бессмысленный и тусклый свет.<br/>\n",
    "Живи еще хоть четверть века -<br/>\n",
    "Все будет так. Исхода нет.<br/>\n",
    "\n",
    "Умрешь - начнешь опять сначала<br/>\n",
    "И повторится все, как встарь:<br/>\n",
    "Ночь, ледяная рябь канала,<br/>\n",
    "Аптека, улица, фонарь.<br/>\n",
    "\n",
    "Пусть каждая строчка - это будет условный документ. Переведем эти документы в числовое представление по методу мешка слов (Bag Of Words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "\n",
    "corpus = \\\n",
    "[u'Ночь, улица, фонарь, аптека,',\n",
    " u'Бессмысленный и тусклый свет.',\n",
    " u'Живи еще хоть четверть века -',\n",
    " u'Все будет так. Исхода нет.',\n",
    " u'Умрешь - начнешь опять сначала',\n",
    " u'И повторится все, как встарь:',\n",
    " u'Ночь, ледяная рябь канала,',\n",
    " u'Аптека, улица, фонарь.']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8L, 27L)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vect = CountVectorizer()\n",
    "\n",
    "bow = vect.fit_transform(corpus)\n",
    "bow = bow.toarray()\n",
    "bow.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Получили матрицу размера $8 \\times 27$, так как у нас 8 строчек и выделилось 27 униальных слов. Элемент матрицы показывает сколько раз слово встречалось в соответствующем документе."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0,\n",
       "        1, 0, 1, 0, 0],\n",
       "       [0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1,\n",
       "        0, 0, 0, 0, 0],\n",
       "       [0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "        0, 0, 0, 1, 1],\n",
       "       [0, 0, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0,\n",
       "        0, 0, 0, 0, 0],\n",
       "       [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0,\n",
       "        0, 1, 0, 0, 0],\n",
       "       [0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0,\n",
       "        0, 0, 0, 0, 0],\n",
       "       [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0,\n",
       "        0, 0, 0, 0, 0],\n",
       "       [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "        1, 0, 1, 0, 0]], dtype=int64)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "столбец 0: аптека\n",
      "столбец 1: бессмысленный\n",
      "столбец 2: будет\n",
      "столбец 3: века\n",
      "столбец 4: все\n",
      "столбец 5: встарь\n",
      "столбец 6: еще\n",
      "столбец 7: живи\n",
      "столбец 8: исхода\n",
      "столбец 9: как\n",
      "столбец 10: канала\n",
      "столбец 11: ледяная\n",
      "столбец 12: начнешь\n",
      "столбец 13: нет\n",
      "столбец 14: ночь\n",
      "столбец 15: опять\n",
      "столбец 16: повторится\n",
      "столбец 17: рябь\n",
      "столбец 18: свет\n",
      "столбец 19: сначала\n",
      "столбец 20: так\n",
      "столбец 21: тусклый\n",
      "столбец 22: улица\n",
      "столбец 23: умрешь\n",
      "столбец 24: фонарь\n",
      "столбец 25: хоть\n",
      "столбец 26: четверть\n"
     ]
    }
   ],
   "source": [
    "# Посмотрим какие слова стоят по столбцам\n",
    "for i, w in enumerate(vect.get_feature_names()):\n",
    "    print u'столбец %d: %s' % (i, w)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Туториал: Имплементация собственного метода"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Для решения задач вида \"имплементируйте метод X\" крайне желательно (но необязательно) реализовывать методы с помощью интерфейса sklearn. Ниже вы увидите пример очень простого **классификатора**. Используйте эти наработи для выполнения задания."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Загрузите [датасет](https://archive.ics.uci.edu/ml/machine-learning-databases/wine-quality/winequality-red.csv) с информацией о характеристиках вина и его качестве. (В метод `pd.read_csv()` можно подать веб-ссылку на данные и пандас их скачает и преобразует в `DataFrame`)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "df_wine = pd.read_csv('https://archive.ics.uci.edu/ml/machine-learning-databases/wine-quality/winequality-red.csv', sep=';')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>fixed acidity</th>\n",
       "      <th>volatile acidity</th>\n",
       "      <th>citric acid</th>\n",
       "      <th>residual sugar</th>\n",
       "      <th>chlorides</th>\n",
       "      <th>free sulfur dioxide</th>\n",
       "      <th>total sulfur dioxide</th>\n",
       "      <th>density</th>\n",
       "      <th>pH</th>\n",
       "      <th>sulphates</th>\n",
       "      <th>alcohol</th>\n",
       "      <th>quality</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>7.4</td>\n",
       "      <td>0.70</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.9</td>\n",
       "      <td>0.076</td>\n",
       "      <td>11.0</td>\n",
       "      <td>34.0</td>\n",
       "      <td>0.9978</td>\n",
       "      <td>3.51</td>\n",
       "      <td>0.56</td>\n",
       "      <td>9.4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>7.8</td>\n",
       "      <td>0.88</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2.6</td>\n",
       "      <td>0.098</td>\n",
       "      <td>25.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>0.9968</td>\n",
       "      <td>3.20</td>\n",
       "      <td>0.68</td>\n",
       "      <td>9.8</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>7.8</td>\n",
       "      <td>0.76</td>\n",
       "      <td>0.04</td>\n",
       "      <td>2.3</td>\n",
       "      <td>0.092</td>\n",
       "      <td>15.0</td>\n",
       "      <td>54.0</td>\n",
       "      <td>0.9970</td>\n",
       "      <td>3.26</td>\n",
       "      <td>0.65</td>\n",
       "      <td>9.8</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11.2</td>\n",
       "      <td>0.28</td>\n",
       "      <td>0.56</td>\n",
       "      <td>1.9</td>\n",
       "      <td>0.075</td>\n",
       "      <td>17.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>0.9980</td>\n",
       "      <td>3.16</td>\n",
       "      <td>0.58</td>\n",
       "      <td>9.8</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7.4</td>\n",
       "      <td>0.70</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.9</td>\n",
       "      <td>0.076</td>\n",
       "      <td>11.0</td>\n",
       "      <td>34.0</td>\n",
       "      <td>0.9978</td>\n",
       "      <td>3.51</td>\n",
       "      <td>0.56</td>\n",
       "      <td>9.4</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   fixed acidity  volatile acidity  citric acid  residual sugar  chlorides  \\\n",
       "0            7.4              0.70         0.00             1.9      0.076   \n",
       "1            7.8              0.88         0.00             2.6      0.098   \n",
       "2            7.8              0.76         0.04             2.3      0.092   \n",
       "3           11.2              0.28         0.56             1.9      0.075   \n",
       "4            7.4              0.70         0.00             1.9      0.076   \n",
       "\n",
       "   free sulfur dioxide  total sulfur dioxide  density    pH  sulphates  \\\n",
       "0                 11.0                  34.0   0.9978  3.51       0.56   \n",
       "1                 25.0                  67.0   0.9968  3.20       0.68   \n",
       "2                 15.0                  54.0   0.9970  3.26       0.65   \n",
       "3                 17.0                  60.0   0.9980  3.16       0.58   \n",
       "4                 11.0                  34.0   0.9978  3.51       0.56   \n",
       "\n",
       "   alcohol  quality  \n",
       "0      9.4        5  \n",
       "1      9.8        5  \n",
       "2      9.8        5  \n",
       "3      9.8        6  \n",
       "4      9.4        5  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_wine.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Наша задача состоит в предсказании целевого признака `quality` по остальным признакам. `quality` - это вещественная переменная (оценка качества от 0 до 10). Упростим себе задачу - создадим новый столбец `quality_cat` по следующим правилам:\n",
    "\n",
    "* `quality_cat = 1` если `quality > 5` \n",
    "* `quality_cat = 0` - иначе \n",
    "\n",
    "Таким образом, `quality_cat` характерует \"плохое\" и \"хорошее\" вино."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def prepocess_wine(df):\n",
    "    df.loc[:, 'quality_cat'] = (df.loc[:, 'quality'] > 5).astype(int)\n",
    "    df = df.drop('quality', axis=1)\n",
    "    \n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>fixed acidity</th>\n",
       "      <th>volatile acidity</th>\n",
       "      <th>citric acid</th>\n",
       "      <th>residual sugar</th>\n",
       "      <th>chlorides</th>\n",
       "      <th>free sulfur dioxide</th>\n",
       "      <th>total sulfur dioxide</th>\n",
       "      <th>density</th>\n",
       "      <th>pH</th>\n",
       "      <th>sulphates</th>\n",
       "      <th>alcohol</th>\n",
       "      <th>quality_cat</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>7.4</td>\n",
       "      <td>0.70</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.9</td>\n",
       "      <td>0.076</td>\n",
       "      <td>11.0</td>\n",
       "      <td>34.0</td>\n",
       "      <td>0.9978</td>\n",
       "      <td>3.51</td>\n",
       "      <td>0.56</td>\n",
       "      <td>9.4</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>7.8</td>\n",
       "      <td>0.88</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2.6</td>\n",
       "      <td>0.098</td>\n",
       "      <td>25.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>0.9968</td>\n",
       "      <td>3.20</td>\n",
       "      <td>0.68</td>\n",
       "      <td>9.8</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>7.8</td>\n",
       "      <td>0.76</td>\n",
       "      <td>0.04</td>\n",
       "      <td>2.3</td>\n",
       "      <td>0.092</td>\n",
       "      <td>15.0</td>\n",
       "      <td>54.0</td>\n",
       "      <td>0.9970</td>\n",
       "      <td>3.26</td>\n",
       "      <td>0.65</td>\n",
       "      <td>9.8</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11.2</td>\n",
       "      <td>0.28</td>\n",
       "      <td>0.56</td>\n",
       "      <td>1.9</td>\n",
       "      <td>0.075</td>\n",
       "      <td>17.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>0.9980</td>\n",
       "      <td>3.16</td>\n",
       "      <td>0.58</td>\n",
       "      <td>9.8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7.4</td>\n",
       "      <td>0.70</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.9</td>\n",
       "      <td>0.076</td>\n",
       "      <td>11.0</td>\n",
       "      <td>34.0</td>\n",
       "      <td>0.9978</td>\n",
       "      <td>3.51</td>\n",
       "      <td>0.56</td>\n",
       "      <td>9.4</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   fixed acidity  volatile acidity  citric acid  residual sugar  chlorides  \\\n",
       "0            7.4              0.70         0.00             1.9      0.076   \n",
       "1            7.8              0.88         0.00             2.6      0.098   \n",
       "2            7.8              0.76         0.04             2.3      0.092   \n",
       "3           11.2              0.28         0.56             1.9      0.075   \n",
       "4            7.4              0.70         0.00             1.9      0.076   \n",
       "\n",
       "   free sulfur dioxide  total sulfur dioxide  density    pH  sulphates  \\\n",
       "0                 11.0                  34.0   0.9978  3.51       0.56   \n",
       "1                 25.0                  67.0   0.9968  3.20       0.68   \n",
       "2                 15.0                  54.0   0.9970  3.26       0.65   \n",
       "3                 17.0                  60.0   0.9980  3.16       0.58   \n",
       "4                 11.0                  34.0   0.9978  3.51       0.56   \n",
       "\n",
       "   alcohol  quality_cat  \n",
       "0      9.4            0  \n",
       "1      9.8            0  \n",
       "2      9.8            0  \n",
       "3      9.8            1  \n",
       "4      9.4            0  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_wine = df_wine.pipe(prepocess_wine)\n",
    "df_wine.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Нарисуем распределение хорошего и плохого вина по признаку `alcohol`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0xa569ac8>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtoAAAFRCAYAAABKcQq1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAwjaAAAAgAH/HxHu6b/6d2n1AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xc330668>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "_, ax = plt.subplots(1,1)\n",
    "df_wine.query('quality_cat == 1').hist('alcohol', alpha=0.3, bins=20, label='good', ax=ax)\n",
    "df_wine.query('quality_cat == 0').hist('alcohol', alpha=0.3, bins=20, label='bad', ax=ax)\n",
    "ax.set_xlabel('alcohol')\n",
    "ax.set_ylabel('count')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Визуально кажется, что можно найти некоторое пороговое значение признака `alcohol` $\\theta$, чтобы хорошо дискриминировать между плохими и хорошим вином:\n",
    "* Если `alcohol >` $\\theta$ - то предсказываем, что вино хорошее\n",
    "* Если `alcohol <=` $\\theta$ - то предсказываем, что вино плохое\n",
    "\n",
    "Обучаеть классификатор будем очень просто: перебирать значения признака alchol и считать **ошибку** классификации. То значение, где ошибка минимальна и будет оптимальным.\n",
    "\n",
    "Разделим целевой признак от остальных признаков"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X, y = df_wine.iloc[:, :-1].values, df_wine.iloc[:, -1].values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Далее вы увидете пример реализации этого простого классификатора."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.base import BaseEstimator, ClassifierMixin\n",
    "\n",
    "class SimpleClassifier(BaseEstimator, ClassifierMixin): \n",
    "    \n",
    "    def __init__(self):\n",
    "        \"\"\"\n",
    "        Инициализация метода\n",
    "        \"\"\"\n",
    "        \n",
    "        self.threshold = 0\n",
    "\n",
    "    def fit(self, X, y=None):\n",
    "        \"\"\"\n",
    "        Выполняет процедуру обучения классификатора. В данном случае - перебор занчения признака alcohol.\n",
    "        Договоримся, что признак alcohol будет в последнем столбце матрицы X\n",
    "        \"\"\"\n",
    "        \n",
    "        alcohol = X[:,-1]\n",
    "        n = float(y.shape[0])\n",
    "        \n",
    "        # Задаем пространство поиска\n",
    "        alco_range = np.arange(alcohol.min(), alcohol.max(), 0.1)\n",
    "        \n",
    "        min_error_rate = 1\n",
    "                \n",
    "        #  Начинаем поиск\n",
    "        for theta in alco_range:\n",
    "            prediction = alcohol > theta\n",
    "            \n",
    "            error_rate = sum(prediction != y)/n\n",
    "            \n",
    "            if error_rate < min_error_rate:\n",
    "                min_error_rate = error_rate\n",
    "                self.threshold = theta  \n",
    "        \n",
    "        return self\n",
    "\n",
    "    def predict(self, X, y=None):\n",
    "        \"\"\"\n",
    "        Процедура предсказания класса по обученному порогу\n",
    "        \"\"\"\n",
    "        \n",
    "        return (X[:,-1] > self.threshold).astype(int)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Обучаем модель"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Оптимальный порог: 10.000000\n"
     ]
    }
   ],
   "source": [
    "model = SimpleClassifier()\n",
    "model.fit(X, y)\n",
    "\n",
    "print 'Оптимальный порог: %f' % model.threshold"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Получаем предсказание модели (на обучающей выборке)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_hat = model.predict(X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Оценим долю ошибок на обучающей выборке"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Доля ошибок на обучающей выборке: 0.295184\n"
     ]
    }
   ],
   "source": [
    "error_rate = sum(y_hat != y)/float(y.shape[0])\n",
    "print 'Доля ошибок на обучающей выборке: %f' % error_rate"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Проверка метода на контрольной выборке\n",
    "\n",
    "Мы обучили наш метод на данных о красном виде. \n",
    "\n",
    "Посмотрим, выполняется ли так же хорошо закономерность между качеством и уровнем алкоголя в белом вине."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>fixed acidity</th>\n",
       "      <th>volatile acidity</th>\n",
       "      <th>citric acid</th>\n",
       "      <th>residual sugar</th>\n",
       "      <th>chlorides</th>\n",
       "      <th>free sulfur dioxide</th>\n",
       "      <th>total sulfur dioxide</th>\n",
       "      <th>density</th>\n",
       "      <th>pH</th>\n",
       "      <th>sulphates</th>\n",
       "      <th>alcohol</th>\n",
       "      <th>quality_cat</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>7.0</td>\n",
       "      <td>0.27</td>\n",
       "      <td>0.36</td>\n",
       "      <td>20.7</td>\n",
       "      <td>0.045</td>\n",
       "      <td>45.0</td>\n",
       "      <td>170.0</td>\n",
       "      <td>1.0010</td>\n",
       "      <td>3.00</td>\n",
       "      <td>0.45</td>\n",
       "      <td>8.8</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>6.3</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.34</td>\n",
       "      <td>1.6</td>\n",
       "      <td>0.049</td>\n",
       "      <td>14.0</td>\n",
       "      <td>132.0</td>\n",
       "      <td>0.9940</td>\n",
       "      <td>3.30</td>\n",
       "      <td>0.49</td>\n",
       "      <td>9.5</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>8.1</td>\n",
       "      <td>0.28</td>\n",
       "      <td>0.40</td>\n",
       "      <td>6.9</td>\n",
       "      <td>0.050</td>\n",
       "      <td>30.0</td>\n",
       "      <td>97.0</td>\n",
       "      <td>0.9951</td>\n",
       "      <td>3.26</td>\n",
       "      <td>0.44</td>\n",
       "      <td>10.1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>7.2</td>\n",
       "      <td>0.23</td>\n",
       "      <td>0.32</td>\n",
       "      <td>8.5</td>\n",
       "      <td>0.058</td>\n",
       "      <td>47.0</td>\n",
       "      <td>186.0</td>\n",
       "      <td>0.9956</td>\n",
       "      <td>3.19</td>\n",
       "      <td>0.40</td>\n",
       "      <td>9.9</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>7.2</td>\n",
       "      <td>0.23</td>\n",
       "      <td>0.32</td>\n",
       "      <td>8.5</td>\n",
       "      <td>0.058</td>\n",
       "      <td>47.0</td>\n",
       "      <td>186.0</td>\n",
       "      <td>0.9956</td>\n",
       "      <td>3.19</td>\n",
       "      <td>0.40</td>\n",
       "      <td>9.9</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   fixed acidity  volatile acidity  citric acid  residual sugar  chlorides  \\\n",
       "0            7.0              0.27         0.36            20.7      0.045   \n",
       "1            6.3              0.30         0.34             1.6      0.049   \n",
       "2            8.1              0.28         0.40             6.9      0.050   \n",
       "3            7.2              0.23         0.32             8.5      0.058   \n",
       "4            7.2              0.23         0.32             8.5      0.058   \n",
       "\n",
       "   free sulfur dioxide  total sulfur dioxide  density    pH  sulphates  \\\n",
       "0                 45.0                 170.0   1.0010  3.00       0.45   \n",
       "1                 14.0                 132.0   0.9940  3.30       0.49   \n",
       "2                 30.0                  97.0   0.9951  3.26       0.44   \n",
       "3                 47.0                 186.0   0.9956  3.19       0.40   \n",
       "4                 47.0                 186.0   0.9956  3.19       0.40   \n",
       "\n",
       "   alcohol  quality_cat  \n",
       "0      8.8            1  \n",
       "1      9.5            1  \n",
       "2     10.1            1  \n",
       "3      9.9            1  \n",
       "4      9.9            1  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_wine_test = pd.read_csv('https://archive.ics.uci.edu/ml/machine-learning-databases/wine-quality/winequality-white.csv',\n",
    "                           sep=';')\n",
    "\n",
    "df_wine_test = df_wine_test.pipe(prepocess_wine)\n",
    "df_wine_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X_test, y_test = df_wine_test.iloc[:, :-1].values, df_wine_test.iloc[:, -1].values\n",
    "y_hat = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Доля ошибок на контрольной выборке: 0.312577\n"
     ]
    }
   ],
   "source": [
    "error_rate = sum(y_hat != y_test)/float(y_test.shape[0])\n",
    "print 'Доля ошибок на контрольной выборке: %f' % error_rate"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Получилось неплохо и это лишь наивный алгоритм на одном признаке! Ну а выводы делайте сами =)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Имплементация K-means\n",
    "\n",
    "Пользуясь наработками выше, имплементируйте метод k-means.  \n",
    "При инициализации необходимо задавать количество кластеров, функцию расстояния между кластерами (для оригинального k-means - евклидово расстояние) и начальное состояние генератора случайных чисел.\n",
    "\n",
    "После обучения, среди атрибутов класса `Kmeans` должны появится\n",
    "* Метки кластеров для объектов\n",
    "* Координаты центройдов кластеров\n",
    "\n",
    "k-means - это алгоритм **кластеризации**, а не классификации, а посему метод `.predict()` в нем фактически не нужен, но он может возвращать метки ближайшего кластера для объектов."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Задание 1 (2 баллов)\n",
    "Имплементируйте метод k-means\n",
    "\n",
    "Теория для выполнения задания 2 и 3 остаётся на самостоятельное изучение. Теории немного и она совсем простая.\n",
    "\n",
    "###### Задание 2 (2 балла)\n",
    "Имплементируйте класс MiniBatchKMeans, который является классом наследником Kmeans.\n",
    "\n",
    "###### Задание 3 (2 балла)\n",
    "Превратите k-means в k-means++. Для этого нужно реализовать метод в классе Kmeans, который будет инициализировать более \"хорошие\" значения центроидов. Чтобы использовался метод k-means++, в параметр init необходимо передать строковое значение 'k-means' (по умолчанию 'random').\n",
    "\n",
    "###### Задание 4 (2 балла)\n",
    "В пункте \"Проверка корректности метода\" нужно нарисовать графики, которые показывают зависимость времени выполнения алгоритма от количества сэмплов. Графики должны быть  нарисованы для различных комбинаций реализаций алгоритма (k-means, k-means++, k-means с MiniBatchKMeans, k-means++ с MiniBatchKMeans). Для этого рекомендуется использовать инструмент %timeit. График достаточно построить на 5-10 точках.\n",
    "\n",
    "###### Задание 5 (2 балла)\n",
    "В пункте \"Применение K-means на реальных данных\" нужно сравнить различные реализации k-means (k-means, k-means++, k-means с MiniBatchKMeans, k-means++ с MiniBatchKMeans). После чего написать вывод, в котором должно быть объяснение того, почему один алгоритм оказался лучше остальных или почему не было выявлено лучшего алгоритма."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from sklearn.base import ClusterMixin\n",
    "from sklearn.neighbors import DistanceMetric\n",
    "from sklearn.base import BaseEstimator, ClassifierMixin\n",
    "class Kmeans(BaseEstimator, ClusterMixin): \n",
    "    \n",
    "    def __init__(self, k=2, metric='euclidean', max_iter=1000, random_state=None, init='random'):\n",
    "        \"\"\"\n",
    "        Инициализация метода\n",
    "        :k - количество кластеров\n",
    "        :metric - функция расстояния между объектами\n",
    "        :max_iter - максиальное количество итераций\n",
    "        :random_state - seed для инициализации генератора случайных чисел\n",
    "        \"\"\"\n",
    "        \n",
    "        self.k = k\n",
    "        self.random_state = random_state\n",
    "        self.metric = metric\n",
    "        self.max_iter = max_iter\n",
    "        self.init = init\n",
    "    \n",
    "    def init_centroids(self, X):\n",
    "        if self.init == 'k-means':\n",
    "            self.centroids = X[np.random.choice(xrange(X.shape[0]), 1)]\n",
    "            # Массив с центройдами кластеров\n",
    "            for i in xrange(self.k - 1):\n",
    "                dist = DistanceMetric.get_metric(self.metric)\n",
    "                res = dist.pairwise(X, self.centroids)\n",
    "                probs = np.amax(res, axis=1)\n",
    "                self.centroids = np.vstack((self.centroids, X[np.random.choice(xrange(X.shape[0]), 1, p=np.divide(probs, np.sum(probs, axis=0)))]))\n",
    "        else:\n",
    "            self.centroids = X[np.random.choice(xrange(X.shape[0]), self.k)]\n",
    "    \n",
    "    def fit(self, X, y=None):\n",
    "        \"\"\"\n",
    "        Процедура обучения k-means\n",
    "        \"\"\"\n",
    "        \n",
    "        # Инициализация генератора случайных чисел\n",
    "        np.random.seed(self.random_state)\n",
    "        \n",
    "        # Массив с метками кластеров для каждого объекта из X\n",
    "        self.labels = np.empty(X.shape[0])\n",
    "        \n",
    "        # Массив с центройдами кластеров\n",
    "        self.init_centroids(X)\n",
    "        \n",
    "        \n",
    "        for n in xrange(self.max_iter):\n",
    "            dist = DistanceMetric.get_metric(self.metric)\n",
    "            res = dist.pairwise(X, self.centroids)\n",
    "            self.labels = res.argmin(axis=1)\n",
    "            cnts = np.zeros((self.k, 1), dtype='int')\n",
    "            for i in xrange(X.shape[0]):\n",
    "                self.centroids[self.labels[i]] += X[i]\n",
    "                cnts[self.labels[i]][0] += 1\n",
    "                \n",
    "            self.centroids = np.divide(self.centroids, cnts)\n",
    "        \n",
    "        \n",
    "        return self\n",
    "\n",
    "    def predict(self, X, y=None):\n",
    "            dist = DistanceMetric.get_metric(self.metric)\n",
    "            res = dist.pairwise(X, self.centroids)\n",
    "            return res.argmin(axis=1)\n",
    "\n",
    "class MiniBatchKMeans(Kmeans):\n",
    "    def __init__(self, k=2, metric='euclidean', max_iter=1000, random_state=None, init='random', batch_size = 10000):\n",
    "        Kmeans.__init__(self, k, metric, max_iter, random_state, init)\n",
    "        self.batch_size = batch_size\n",
    "        \n",
    "    def fit(self, X, y=None):\n",
    "        \"\"\"\n",
    "        Процедура обучения k-means-minibatch\n",
    "        \"\"\"\n",
    "        \n",
    "        # Инициализация генератора случайных чисел\n",
    "        np.random.seed(self.random_state)\n",
    "        \n",
    "        # Массив с метками кластеров для рандомных объектов из X\n",
    "        self.labels = np.empty(min(X.shape[0], self.batch_size))\n",
    "        \n",
    "        # Массив с центройдами кластеров\n",
    "        self.init_centroids(X);\n",
    "        \n",
    "        \n",
    "        self.cnts = np.zeros(self.k, dtype='int')\n",
    "        \n",
    "        \n",
    "        for n in xrange(self.max_iter):\n",
    "            X1 = X[np.random.choice(xrange(X.shape[0]), min(X.shape[0], self.batch_size))]\n",
    "            \n",
    "            dist = DistanceMetric.get_metric(self.metric)\n",
    "            res = dist.pairwise(X1, self.centroids)\n",
    "            self.labels = res.argmin(axis=1)\n",
    "            \n",
    "            for i in xrange(X1.shape[0]):\n",
    "                self.cnts[self.labels[i]] += 1\n",
    "                n = 1./self.cnts[self.labels[i]]\n",
    "                self.centroids[self.labels[i]] = (1 - n) * self.centroids[self.labels[i]] + n * X1[i]\n",
    "            \n",
    "        return self\n",
    "        "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Проверка корректности метода\n",
    "\n",
    "Перед тем как применять алгоритм на реальных данных, нужно испытать его на простых \"игрушечных\" данных.\n",
    "\n",
    "Если алгоритм реализован правильно, то метод должен идеально разбивать на 3 кластера данные ниже. Проверьте это.\n",
    "\n",
    "ВНИМАНИЕ! Проверка должна быть осуществлена на всех реализациях, иначе реализация не будет зачтена!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.datasets import make_blobs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "X, y = make_blobs(n_samples=100, n_features=2, centers=3, cluster_std=1, \n",
    "           center_box=(-10.0, 10.0), shuffle=False, random_state=1234)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0xcd34c50>"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsUAAAEyCAYAAAAStK0kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIhIzFNTLCIiIiIx7/8Bk0tZKq1JjzAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xcb1a940>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(X[:,0], X[:, 1], c=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[12, 14],\n",
       "       [16, 18],\n",
       "       [19, 20]])"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "n = np.array([[12, 14], [16, 18]])\n",
    "np.vstack((n, [[19, 20]]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Проверьте Вашу имплементацию на простых данных (без этого пункта задание не считается выполненным).   \n",
    "КОММЕНТАРИИ НЕ СТИРАТЬ!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Rnd_state = 1000;"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-1.29992907  5.919722  ]\n",
      " [ 5.81037396 -4.9431358 ]\n",
      " [-5.99382611  2.25004374]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsUAAAEyCAYAAAAStK0kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3Xl8lNXZ//HPmZlM9j1sYQdBFllEVMRd0CpSnz7ua23V\nWpdWH6tW6/pTa8WqrbW1LtW6d7EuVStuKa6Aiiiy7zsBQvZkJrOf3x+BaJwASWaSSZjv+x9fc+bc\n575yGPCak3Nfx1hrLSIiIiIiScyR6ABERERERBJNSbGIiIiIJD0lxSIiIiKS9JQUi4iIiEjSU1Is\nIiIiIklPSbGIiIiIJD0lxSIiIiKS9JQUi4iIiEjSU1IsIiIiIklPSbGIiIiIJD1Xom5cWlra5muK\nioooLy/vgGj2fZq72Gj+2k9zFxvNX/tp7mKj+Ws/zV1s4j1/xcXFreqnlWIRERERSXpKikVEREQk\n6cVl+8R//vMfZs2ahTGG/v37c8UVV+B2u+MxtIiIiIhIh4t5pbiyspK33nqLGTNm8MADDxCJRJgz\nZ048YhMRERER6RRx2T4RiUQIBAKEw2ECgQD5+fnxGFZEREREpFMYa62NdZCZM2fy97//Hbfbzbhx\n47jqqqui+pSUlFBSUgLAjBkzCAQCbb6Py+UiFArFGm5S0tzFRvPXfpq72Gj+2k9zFxvNX/tp7mIT\n7/lr7ZbemJPi+vp6HnjgAa655hoyMjL43e9+x6RJkzjqqKP2eJ1KsnUuzV1sNH/tp7mLjeav/TR3\nsdH8tZ/mLjaJKskW84N2ixYtomfPnuTk5ABw6KGHsnLlyr0mxSLS/ZWu2c7rv3+H+mov/Uf24ZSr\nv0d6VlqiwxIREWmzmJPioqIiVq1ahd/vx+12s2jRIoYOHRqP2ESShrWW9Ys2Ub2tlv0nDSUjJz3R\nIe3V6i/W8fAVT1NZWg3Aog+WsWz2Kn718lWkpqv6jIiIdC8xJ8XDhg1j0qRJ3HDDDTidTgYNGsTU\nqVPjEZtIUqiv8vDgRX9h07JSfPV+evQvYMqPjuSknx6X6ND26OX7ZzYlxLusW7SJ9/76EdOvTN5/\nAwK+IO8++SFr5q8nt2c2p1z1PQqK8xIdloiI7EVc6hSfeeaZnHnmmfEYSiTpPHnd31k1b13T6x2b\nKpn56CwmnDCGXoN7JDCyPaurqI9utLB+4cbOD6aLCPpD/Pbsh1k1fx3sfFpj8YfL+dljFzFobP/E\nBiciInukE+1EEmzLym1RbbU76njv6Y8SEE3rZeZltNjeZ2ivTo6k65j1/GzWfLW+KSGGxi85/7r3\njYTFJCIiraOkWCTBjGm5PcUdl1/kdJiTLjuOnB7Zzdr6Du/NSZd17W0fHWnV52uJhKML+lRvr01A\nNCIi0hZd+/+6Iklg8LgBbFu7o1lbfp88jr/o6ARF1Drjp4zmiod/yJuPzMJX56OofyFn3fz9bvGQ\nYEfpOaiwxfZknhMRke5CSbFIgv343rNoqPOxbuEm/B4/BX3z+f7PjqegT9d/OGvk5OGMnDw80WF0\nGdMum8KXby9i65qyprasgkym/OjIBEYlIiKtoaRYJMFSM1K55ulLqd5eQ21FPX2H98bpciY6LGmH\nrPxMfvHcT/nHna9RUVpFWlYqx//oKCZOG5fo0EREZC+UFIt0EXm9csnrlZvoMCRGPQcUcdUTFyc6\nDBERaSM9aCciIiIiSU8rxSKStJZ8vIIP/jYXh8Mw9UdHMuzgIVF9QsEw9VUesgsyta1FRGQfpqRY\nRJLSi795nVnPzaahzgfAog+XM+3y45h+5fFNff79u7f59PUv8dZ4ySrI5JhzJnPCJV27KoiIiLSP\ntk+ISNKpr/Iw97UvmxJiAE+1l4/+8Sl+rx+Aj1/8jJmPzWLr6u3U7Khjy4pt/Pv3b7H4o+WJCltE\nRDqQkmKRbiwSjvD1f5fwyb8+x1PtTXQ43ca6hRup3FIV1V5ZWk3pqu0AzH55Hn6Pv9n7npoGSp7+\nuFNiFBGRzqXtEyLd1LZ1Zfzp0qfYuno7oWCYon4FnPjTYzn+x0clOrQur+eAIrIKMqmv9DRrz8rP\npKC4sT50OBhu8dpQINTh8YmISOfTSrFIN/XX6//BpmWlhHYmb+WbK5n55xJqduhI4b3pNbgHQw8c\n2KzN4TQMP2QIuT1yABjynfcBHC4HBxw1olNiFBGRzqWkWKQbaqj3sWNDRVR75dYaPnlpXgIi6n5+\n/vjFTLnwCAaN7c/gcQM44ZJj+OlDFzS9f9r1JzPqiOGkZaUCjUc1jz9uFMdfpJV4EZF9kbZPiHRD\nTqcDh7OF77QG0ncmcbJnKakufnj3Gbt9352Wwi//fgUrPlvLmq/WM/Kw/RgyPnr1+Ns8NV5WzVtL\nQXE+A0b1jXfIIiLSgZQUi3RD7nQ3Aw/oR/nmymbtvQb14PDTDk5QVPseYwwjJg1lxKShe+37xp/e\n44Pn51C+uZL0nHQGjCrm//76EzJy0jshUhERiZW2T4h0Uz958DzGTRlNQZ9csgoyGTSmPxfffw6p\nGVop7mybV2zlncffb/qS0lDbwIpP1/D0jf9McGQiItJaWikW6abSs9L4xTOXUl/lwefxU9g3H2NM\nosNKSu/99SPqvlPJAmDj0tIERCMiIu2hpFikm8vKzyQrPzPRYcRFJBwhFAzjTktJdCht4kxp+Zdu\nDoe+pIiIdBdKikUk4SLhCC/c/gqLP15BoCFAfp88zrr5FPY/ZO97ebuCky49jvlvLaR6e/NyeHt7\nME9ERLoOJcUiknD/uPt13n9+NuFQBGg8We4v//cC/+/Na7v0Krjf6+el377J+oWbdj5QZ/DWNZCe\nlcaQ8QP44d2nJzpEERFpJSXFIpJwSz5a3pQQ77JjYwXvPvkBp153coKi2rNIJML95z/Kys/XNrWl\nZ6cx5YdHcOJPjiGvV24CoxMRkbZSUiwiCRfwBVtsr9xag6faS3pOGg5H/Ivl+L1+Xr5vJusXbcad\nlsLUHx/J+CmjW3XtgpIlrFu4qVlbQ52PxR8u56ybT4l7rCIi0rGUFItIwvUcWETZ+vJmbSmpLpbP\nXc1NU2aQlZ/B0ecexgkXHR23e0bCEe479xFWfbGuqW3tgg2cceP3Ofb8yXu9ftW8dQRbSOY91V4C\nviCp6e64xSoiIh1PdYpFpMN5ary888QHvPXof6mrrI96/8LfnEHf/XvjcDZWa0jNcBMJW3ZsrKB6\new2bl2/l3w+8zcL3l8Ytpi9mfh210uup9vL+87Nbdf3oI4aTmhGd+GYVZHW76hkiIqKVYpFur7a8\njtcefIfyzZX02a8X3//58WTmZiQ6rCZz35jPY9c+w46NFQC899THnHrdNI4445CmPj0HFnHHzOuZ\n88o8yjaUs2zOKtZ8uaHZOJ4aLyXPfMLYY0e16r4LP1jK+8/NJhy2HDr9QCafNrFZHecVn68hFAhF\nXVdf5SEcCuN0Ofc4/uij9mfYxMEs/ngF2Ma2rIJMpv7oCNWLFhHphpQUi3RjOzZXcP+5j7Bt7Q6g\ncZ/rog+W8at//bxLVG0Ih8I8e9s/mxJigIotVbz+0LscMn087m9tMUhJdXH0OYcB8Jsz/tjieCF/\ny3uPv+u1B9/mrcc/oKG2AYCln6xk+aerufj+c5r6jJw8jA///mnUFojsgsy9JsTQeAT0NU9fypuP\nzGLlZ6txp6dwwsXHMHLysFbFKCIiXUtckmKPx8Ojjz7Kpk2bMMZw+eWXM3z48HgMLSJ78K97/tOU\nEO+yeflWXv3dW1xwV+LLgW1evpXtG3ZEtZdtKGfVF+sYfeT+LV6334EDWTF3dfNGA/sfuve6xYGG\nALNf/qIpIQYI+oJ8PWspFaVVFBbnAzDhe2MYMn4gKz795j6NK71HNhtv1bx1vPX4LPzeAMMPHsy0\ny6eSktr4T6fL7eJ/rj4BOGGvcYmISNcWl6T4qaeeYvz48Vx77bWEQiH8fn88hhWRvagsrWqx/buJ\ncqJk5KSTmp6K3xto1u5Od+9xJfsHvziJtV9vZO1XG/B7A6RlpjJs4mBOvmLqXu9ZtrGCmh21Ue01\nZbWsnr++KSl2OBxc99xPef2P77Lmyw2409187ydHM2ryN1/o5776BX+741Vqyxv3QS/+aDnL5q7m\nl3+/okOqYYiISOLEnBR7vV6WLVvGlVde2Tigy4XLpV0ZIp1hd3uHc4qyOjmSlvUYUMjA0f1Y9NGy\nZu19h/dmwOi+u73OnZbCDf+4kiUfr2DVF+sYedgw9p80tFV7dQv65JGVn4mvvvmX86z8TPqP7NP0\n2lPjZdW8tRwy/UBO/+X0Fsd654kPmhJiACysnr+er95bzEHfG7vXWOKlvsrDkk9WUlicx9AJg7Rn\nWUSkAxhrrY1lgPXr1/PYY4/Rr18/NmzYwJAhQ/jRj35EWlpas34lJSWUlJQAMGPGDAKBQEvD7ZHL\n5SIUin4wRvZOcxebrjp/Kz5fzd3nPEjFlm9WjHsNLOLut2+iz+BeCYzsG96aBn574Z/YsHQz1lr6\nDe/DL564jII++R12z4d//lfee/ZDgv7GPzPjMBxy0oHc9vK1APzrvtd568lZbF+/g4ycdIaMG8it\n//oFWXnfrF6HwxEuGXkNZRvLo8affvnxXP77H3VY/N/20v1v8ObjJZRtLCctK42Bo/tx+8vXkVuU\n3Sn378666t/b7kLz136au9jEe/7c7taVyIw5KV6zZg0333wzd911F8OGDeOpp54iPT2ds88+e4/X\nlZaWtvleRUVFlJdH/w9K9k5zF5uuPH8r563l9T+8g6faS05RNqfdMJ0BI4sTHVaTXXMXaAhgoVPq\n91prmfnILBb8dwk2HGHYwUM47fppuNwutqzcyj1n/Im6iual4Q6ZPp4rH/1xs7Zbv3cfG5dsbtbm\ncju55HfnMWhMP3ZsrGDI+IEd9lDj1jXbuef0P0VtB5nwvTFc/eQlHXLPfUlX/nvbHWj+2k9zF5t4\nz19xcev+nxjzPofCwkIKCwsZNqzxietJkybx73//O9ZhRaSVhh88hOuevzzRYeyVuxMPszDGcPIV\nUzj5iilR773314+iEmKATcu3RrUdc+4kXvrtm3hrvnlob8Covnz84qc8d8smPNUNFPYr4IjTD+bU\n66bF94cA3nvqoxb3R29ZuS3u9xIRSXYxJ8V5eXkUFhZSWlpKcXExixYtol+/fvGITUQk7pwpLZdb\na2mb7pQLjyS3Rw6znptN0B+k/8hivHU+5r7yRVOfis2VlDz1EWOPHcV+Bw0CGleq6yrqSc9Ob6pU\n0R7utJa/SDiceshPRCTe4vJE3EUXXcRDDz1EKBSiZ8+eXHHFFfEYVkQk7k689Fi+mPk11dubr8AO\nGT+wxf4Tp41j4rRxTa9vOf7eqD6emgb++8zH7HfQIOa9uYA3/vge1dtrSMtK5YCjRnD+Xae1q1rF\nCRcfzbz/LKB8c2Wz9qEHthyriIi0X1yS4kGDBjFjxox4DCUiskeRSIRVn68l4A8xYtJ+bV6J7dG/\nkNOuP5m3Hp1FRWkV6dlpDB43gAt/c0arrjeOlis/OFwOyjaU88Ltr1C1rQaAmh11VGyeS3pWGmf8\n6vttihMaK2lc/JtzeeHul6nYWk1ahpvB41sfq4iItJ5qp4lIt7F5+VYevepZtq4pIxIM03NwD866\n+RQmnDCmTeMcdfYkJp86kdLV28kpzCKvV26rrx02cTAbl25pOtoZGk/BO+Hio5n56KymhHiXUDDM\noo9WtCspBjjqzMMYcfRQtq0tI6sgi5zCrlFuT0RkX6OkWES6hPLNlfi9Afrs17PFrQbWWp687m9s\nWvpN5Zpta8r4x12vccCR+7f5QT6X28WAUbuvlbw7597+v9TsqGP1/PV4a7wUFOdx3AVHMHB0Pxrq\nGlq8JhyIrbSQw+mgeFjvmMYQEZE9U1IsIglVW1HPw5c9xZaV2wgFQhT1K+D8O09lxGHDmvWr2FJF\n2caKqOvLNpSz8P2lTJw2vlPidbld/Pzxi6guq6VqWw19h/VqSsiPPONQFry3BJ+n+cEhfYZ1jZrR\nIiKye3qEWUQS6tGfP8vyuaupq6inoc7HpmWlPHXDPwk0ND/gx+F04GhhP69xGFzulBbHttay5JOV\nvPPEh2xdWxbXuPN65jB4bP9mK9Sjj9qfST+YQHZBY93ilFRXm/Yri4hI4milWEQ6RG15Hf6GAIWF\nhbvtU1/loXRVdM3d7evLmffm1xx++sFNbQV98ug9tGfzY5eB3oN7cMDRI6LGaKjz8cAPH2PDks0E\nvAFe/0MG6dnpZOVnkJqRytHnHMbkUyfG8BNGM8bw43vP5sSfHseXby+keFhvxk0Z1a7KEyIi0rmU\nFItIq5RvrmRByRL67NeLUYcPw7RU2JfGRPeRK59l84pSQsEwvQf15Nw7/rfFMmKRiMVGog/VtBFL\nKBi9D/fyP/6QP1/5LFvXbCccDNNjQCEX/uYMXC3UHn7+9pdZNW/tt+LyUl/lZcfOLRgbl26hoa6B\nKRce2eo5aK0+Q3py8hVT4z6uiIh0HCXFIrJXz936Ep//ZwG1O+pwp6fQb0Qx1z13GZl5GVF9H/35\nsyz+aHnT69WV6/jLNc9z1zs3RJVPyynMoufAoqiawT0GFHLI9AOjxi4ozueWV6+mbEM5QX+I4mG9\ndpucf/uBvJZ4axr46J+fdUhSLCIi3Y9+pycie7R09ko++dfn1O6oAyDQEGTtVxu4/oi7WPSt5BfA\nU+1l84ro45K3r9vB/Le/bnH8Sx44l4EH9CMlzYVxGHoN7sFpN5xMenZas35Bf4gPXpjD3+54ldqK\nevoO773bhBhocfX4u7w13r32ERGR5KCVYhHZo4/+8Sm+en9Uu6fay1PX/4Nb/v1/FPTJAyAUDBEJ\nR6L6RsKWhhbGAOg1uAf/b+a1rJq3joY6H6MOHxZVXq2itIrf/fAxSldtIxK2fPTPzxg3ZRSXPXTB\nbhPjcVNGsWHJZkKB8G5/tpweObt9T0REkotWikVkj1Izdl//t2JLFTMfndX0OrdHDj36Rz9YV9S/\ngEOm775kmsPhYP9DhzJ+6ugW6w0/e9O/2Lx8K5Fw4/7jhtoGvnxnEQtnLd3tmKdc/T2OOe9weg0q\nIrswK+rnyO+dyylXnbDb60VEJLlopVhE9uiknx7HgveWUF1W2+L7dRXNq0FcdN/Z/PmKZ9i6poxw\nMEzvwT2Z/rMpZOZG7z9urbKN5VFtAW+AOa/OZ9yU0S1eY4zhgrtOI+A7hfoqD1l5Gcx6bjbL5qwi\nIyed6T8/nr46EENERHZSUiwie9R7SE/OuuV/eOamF6O2UaRmuJuVTQPoO7wPd759PYs+WEZDnZ8p\nZx+F1++JKQZ3asur1TlFez/y2J2W0rS948RLj+XES4+NKRYREdk3KSmWuDDUkc3DuMw6LG589jga\n+N9EhyVxMvnUiQw/dAj3n/8I29fuIBK2ZOZmMP740YxpoUaw0+Vk/NQDAMjITo85KR5//Gi2rNxK\n0P9NmbaifvlMu2xKTOOKiIjsoqRY4iBMvvklbrOsqSWFVThsLR4uTGBcEk9FfQv49bs38ulr89m6\najsHTx/PoDH9O+XeP7jmRIL+EAveW4zP4ye/dy6n/fJk8nvndsr9W+Kp9rJhyWZ69C+kx4DdH1Ai\nIiLdg5JiiVkqH5HC6mZtDuMjjQ/w2AvQ85z7DleKkyNOP6TT72uM4cxffZ8zbpxOOBRpVbm1jvTS\nb99k7itfUL65kuyCTIYeNJifPfrjqDrMIiLSfShbkZilsBJjglHthnoMLZfhEmkPY0zCE+Jlc1bx\n36c/onxzJQB1lR4WlCzm73e+mtC4REQkNkqKJWZ+DiNsoysLRMjHktbCFSLd16znZ+Ot9TVvtLDm\nqw2JCUhEROJCv+uTmAUZQ5AJOOxcjGk8KCFsC/DYs4Hdnzgmkigf/uNTZr/0OUFfkOJhvTnnth+Q\nlZ/ZqmsduzksRJ90EZHuTUmxxIGh2t5BOm+RyhwiNgMP5xBmaKIDE4ky89FZvPaHd/DVNa72rl2w\nkS0rt3LLv69p1daMqT86kkUfLsNT3dDUZhyGYYcM6bCYRUSk4ykpljhx0sB0Guz0RAcislvWWua8\nMq8pId5l49JSPnvty6iayy0ZdvAQTrz0OD5+8TMqt1SRVZDJsIlDOOvm/4lbnBsWb+a/z35Caoab\nc284DZMet6FFRGQ3lBSLSNKIhCN4axqi2sPBMOsWbmxVUgxwylUncMLFR7NlxVYKivPjWhrupd++\nyfvPfUJ9lReAL99exJk3fZ9DT5kQt3uIiEg0PWgnIknD6XKS2yM7qj01w83Y40a1aay0zFSGThgU\n14S4uqyWT178rCkhBijfXMnrf3yPSDgSt/uIiEg0JcUiklROvmIquT1zml47XA6GHzK0xZP5OtvC\n95dSta0mqr1qa3VTCTgREekY2j4hIkll4rRx9Brcg/88XIKv3scBR43guB8ejtlNVYnO1KN/Ie4M\nNwFvoFl7WlYqmXnRZQ9FRCR+lBSLSNLpP7KYy//0w0SHEWXEYfvRf/8+zWoeGwP7HTSYzFwlxSIi\nHUlJsYhIF2GM4ZpnLuWpX/6T0lXbcDgdjD1qFKffdHKiQxMR2ecpKRYR6UKyC7K46omLm14XFRVR\nXl6ewIhERJJD3JLiSCTCjTfeSEFBATfeeGO8hpWEipDGO6SZT7A2DQ9nE2JYooMS6RYikQgzH5nF\nwveXYi0ccORwvv/zE3A49XyziEhXFLekeObMmfTt25eGhugaoNI95Zo7SWM2xgTBgNt+SZ29HB8n\nJDo0kS7vr9f9gzmvfkE42Hj0+Zr56yhdvZ3L/3RhgiMTEZGWxGXJoqKigi+//JIpU6bEYzjpAlws\nJ5UvGhPinZymigzzImATF5hIF/Lxi5/ym9Me4vZp9/HwZU9RXVYLQG15HYs/XtGUEAOEQxGWzVlN\nZWl1osIVEZE9iMtK8dNPP83555+/x1XikpISSkpKAJgxYwZFRUVtvo/L5WrXddL2uTPehTh89VHt\nKY4ainLd4IjfgQXdgT577bevzt0bj7zL3+54DW9N40Eb6xdupmxDJb/76A7K11RRVxH996emrJaG\nSj9FY1s/H/vq/HUGzV1sNH/tp7mLTaLmL+akeP78+eTm5jJkyBCWLFmy235Tp05l6tSpTa/b8+CI\nHjhpv7bOXSp9yTVuHKZ5vdRQJIOKygYg2PKF+yh99tpvX527t5+a1ZQQ77JhySZeevB1jjjzEPJ6\n5VC+qfmBG/l98sjundGm+dhX568zaO5io/lrP81dbOI9f8XFxa3qF/P2iRUrVvDFF19w5ZVX8uCD\nD7J48WIeeuihWIeVBPNzWNRDdRGbht8egYqWiIC3Nvo3YzZi2bBkM5m5GUw8aRzuDHfTeylpKUw4\n4QByiqKPmRYRkcSLObs599xzOffccwFYsmQJb7zxBldddVXMgUmiOaiy95JtH8Zl1gJuGuzRNHBa\nogMT6RLyeuZQtr75SkZKqovxUw8A4JzbfsCQ8QOY/fI8rIVJ/zOByadOTESoIiLSClryk92yZFHL\nDXquTqQF/3vtSTz+fy9QtbXxwTmH0zD8kCFMnDa2qc+hp0zg0FMmJCpEERFpg7gmxaNHj2b06NHx\nHFJEpEsadfhwfvWvn/HGH9/DU+Vl1JHDOfb8w3E4VIdYRKQ70kpxtxAii8dxm68xRAjYEdTxMyA1\n0YG1iZs5ZJiZgMFrTybApESHJBKTXoN6cMkD5yY6DBERiQMlxd1ArrmHNN7HmAgALlbhZDvV9rcJ\njqz1sniUDPM6DtP4tL6b+fjtITvrIEfw2aPx8T3AJDROERERSU5Kirs4QzVuFjYlxADGgNsuw8kG\nwgxMYHStY6gjzXzQlBADOIyXND7EmMYNy26+wG2/btzDLCIiItLJtPmti3OyA0NdVLvD1OFkcwIi\najsX63BSFtW+KyEGcJggqeYzHGzrzNBEREREACXFXV6IAUSIPtUlZHsSpHs81BimDxHy9trPaSpx\ns7gTIhIRERFpTklxl5eK104nbHOaWiI2E589DtuKRLMriNCDAGOx9puPm22hzFvE5hBkv06MTERE\nRKSR9hR3A17OJmDHkWFfxpgwXjudIAe1YQQLwXnk8CJheuHlNCyde6pWjb2FMH/FzQLAgA3gYj0O\n03hctLWGAKMJM6hT4xIREREBJcXdRoiR1HJLuw7SyOEenHWzyXB4AEi371FtbyfE8DhHuScu6rn0\nW/FbMuyLpDJ7Z5m5MdRzUSfGIyIiIvINJcX7uBQWkWZmY/A0tbnMFrJ5mCr7hwRGZvByFl57VgJj\nEBEREWmkPcX7uDTexWE8Ue2N1SDaf36zk62ksLBZsi0iIiLSXWmlOI4cVAFBIvRMdChNUliFtY21\njb/Nkkb7Dsrwk2f+Hyksw0ENYXrTYE/Ew4XxCFdEREQkIZQUx4GhmjxzJy42AGHC9KHGXks4wZUU\nDNU4THl0QmwhYMe2a8xs/kQqc5vGdLGVTF7Gbw8hxMgYIxYRERFJDG2fiIM8cyep5kucpgKnqcZt\nlpFn7gZCCY3LRSkOaqLajQE/R7RrTLdZFpVkO0wtGbzSrvFEREREugKtFMfIQQUu1ke1u9iEm3kE\nOGyvY7iZT6Z5Dgc1RMjFY88nwMSYYwvRnwiFUafEhW0PQgz9Tm8LBIEU2retQkRERKT7UlIcI0MA\nCLfwThiDb6/Xu1hBrvkNTlPxrbZNVNlfx7wdwZKNzx5FBq/jMI2xRKwbnz2MCAVN/TJ4iXTzFoY6\nIuThtafi48QWxwzY0bhY3Wy1OGKz8XJqTLGKiIiIJJKS4hiF6U2Y3ji/s00hTB8CHLrX67PMc80S\nYgCnqSCLF6i2v445vnquIGhHkOt+n0AggM8ei4/vNb2fygdkmadxmPqdLWU4eZSQ7UuIMVHj1XEl\nTspJsctwUI3FTYQc0vkv9fTr9ENBREREROJBe4pjZqi1vyBoB2GtE2shZIuptxdhyWjF1S2XNDPU\nt9jeHn6OI5L9CNX23p0rwN8s82aYN76VEDdymmqyzIu7Gc1Ntb2bGvsLIuTiMD5cZguZjpcoMNdg\n8MYtbhHr5SZeAAAgAElEQVQREZHOopXiOAixPxX2Cdx8ioMG/BzeqoQYIGQHkGq+aqG9f7zDbFHj\n9o+W2ve89SPTvIbTVDVrSzGrybD/wKOT6URERKSb0Upx3LgIcAQ+jm91QgxQzyUE7X7YnedoWAtB\nO5R6frKbKyxOSnFQsZv32yZoo8vGWWsI2HF7vM5BdYvtKWZtXOISERER6UxaKU4wSzaV9o9k8Aou\nVuCwVRjjJ99cT9COpI7LgVQAXCwlxzyIi21YXIQYRLW9DUteu+/v4WzcdiFONuAwISI2lSAH4GHP\nxy9HyGmxPWT7tTsWERERkURJsqQ4DDjbeW2AdN7EbRYSsoPwcjqWzLhEZUnHw3nk8mvcZgnGRABI\nYQVOtlFtZwABcs29pJgNTdc5qSSPu6iyD7TjrhFyuB+3mY+DGiw5+CN98HDBzgcE91yWzWPPw8X6\nZg8JBu1APJzbjlhEREREEispkuJU3ifT/BMHVViyabAn4uX0Nozgo8BcSwrLMCaCBdJ4nyp7PxGK\n2hSLoRYXmwjRt9kKr6EaN183JcTQeMiG2y7ByWZcrMbF5qjxXGzAUN3m1eJMniPdvIsxuw4Y8WHw\nY62L1tQpDjCBKnsHWbyAg3pCtjf1/Jh0ZuI2i4jYPOq5gAi92xSXiIiISCLs80mxixXkmIe+9VDY\ndpw8Rdjm4Wdqq8bI5J+ksKSpNq8xkMJ6svkzNfa2VkZiyeYhUs0cnJQTphC/PZg6rgMMTnZgqI26\nymHqcNpNGPzsvh5yGNvKKHZJNZ99KyHedS8PmbxGtW3dwSEhDqDa3tP0qsD8ghQWYYwFA277BdX2\nNkKMbmN0IiIiIp1rn3/QLtP8PapKgsN4yDAzWz1GSgtHGwM42drqMdJ4h3QzE5fZjjFhXKaMDPMu\nGbwM7Dp9LnrVOWR7EmQkfo4gTJ+o98P0IUJhq+NojOVNXKzZzbuR3bTvWTpv7/zi8E167jLbyTZP\ntms8ERERkc60zyfFhobdtPtbPUbEtvxQmSW91WOkmVk4TPN7GhMk1cze1QOvnUbEfnP4RcRm4LPH\nYMnDkkm9vYiQLcZag7UpBO0Qauz1e7irBZqvBrtYSrZ5PCqWxvul0mCPb/XP9G1u8yXGRK9kx6tK\nhkhns9ay4rM1PH/7K3zwwhyC/tDeLxIRkW5rn98+EbTjSOXzZiuYACE7qNVjeLiAVLsApylragvb\nXDy29fuSTSs2OHg5j4AdRyavAGG8djpBDm5638fx+O3huPkMSwYBDqLlP0JLFo+Raj7F0ECEHhD8\nJTCATPMiTlMTfYV10WBPws/Rrf6Zvi1kB7a4FdmS1a7xRBLJWsufr3iahe8vw1fvx+F08N7TH3Pt\nsz+loE/7q72IiEjXtc+vFHs4gwATiNjGsmbWugjYEdRxWavHCNOfKnsrfnsgQTsIvz2AWns1ASa3\negyfPZqITWnWZq0Lv21+FHSIA6ixt1Fj72iWEDddQwZ+jt1ZIaLl7zRZPEmmeZkUsx6X2Y7bLMbp\nuQkHVbtdIQ8yhDr+j9Y8ZNcSL6cRtAObtUVsDl77/XaNJ5JI899ayFfvLcFX3/j3JRKOsHlZKc/d\n+nKCIxMRkY4S80pxeXk5Dz/8MNXV1RhjmDp1KtOmTYtHbHGSQpW9n1Tm4rbzCDISH8fR1h89xBiq\n7O/bHUUDJ+OyK0jjUxzsIEIRfnsQXs5s95i7k2o+xZhgszYT2UwGL+K3h+5cOW++1SFoR8R0T0sW\nVfY+svkTLrYRIQOvPQU/x7VrPAfbSeMDwvTCzxEkwS81pAv5/D9fEfQFo9rL1u1IQDQiItIZYs40\nnE4nF1xwAUOGDKGhoYEbb7yRsWPH0q9fVzrEweBnMv42rOx2RAx1XIvHVuJkI2H6tbmcW+vv1PJq\nsNNUUG8vIZX5uO2XOIyHiE0lxHDq27ByvjsRelJj74x5nCz+Qrp5C6epxFoXIQZSZX9DhF4xjy3S\nGln5LdcgT0nTlzMRkX1VzP/C5+fnk5+fD0B6ejp9+/alsrKyiyXFXUeEAiIUdNj4DsoAX1S7JQOv\nPQFwUm3vwsUyUu08guxPgENo77aJeHOyjnTzn6Z9z8aESGENudxHlb0/wdFJsjj5yqksKFlCxZZv\nKte401OYeNKejz8XEZHuy1hr21ridrfKysq4/fbbeeCBB8jIyGj2XklJCSUlJQDMmDGDQCDQ5vFd\nLhehUJI+AW53VpIwKbvvE1qOs/5qTGRL4yU0prqWLEzqFEIZv6bF2nLxFFyMCX2BdY2HlPFtvtx4\n7sHpfyGq3Zp+hPPe6vj4dyOpP3sx6q5z99V/F/H8nS9RUVpFRnY6k06ZyAW3n47p5M9gd52/rkBz\nFxvNX/tp7mIT7/lzu92t6he3pNjn83H77bdz6qmncuihh+61f2lpaZvvUVRURHl5eXvC68YiZPFn\nUs28nZUkelJrLyPEAVE9880vSTWfN2uzFjz2VNKK7uzguQuRZ27DzUIcpp6IzSTAaKrt3cAeEvnv\nyOQZsh1PRbUH7QAq7LNxjLdtkvOzFx/dfe7CoTAOp6PTk+Fduvv8JZLmLjaav/bT3MUm3vNXXFzc\nqn5xqT4RCoV44IEHOPLII1uVEEvrZfE4mebfpJgNuEwZbrOYPPMbDNFl1VqqCWwMuEzH/8XM5J+k\n8ikOU98Yi/GQyjwyeaZN43j5X0K2+SEl1hoCdmzcYhVpC6fLmbCEWEREOk/MSbG1lkcffZS+ffsy\nffr0eMQk35JqPo86jtllSptOwvu2yG5qAn83yewIjYd3ND8NzxiL2yxu0ziWHGrt9QTs/oRtASHb\nhwZ7PHVcHc9wRURERJqJ+UG7FStW8NFHHzFgwACuv77xdLVzzjmHCRMmxByc7KGSBNGlobz2NFxs\nwGmqm9pCtj8eziGtwyJsZHe7RaLtH7EAE6i0j+Kgmgjp0OHRi4iISLKLOSkeMWIEL774YjxikRaE\nKcbFlmZtEZtJAydF9fVzFDXWSSYvY/AQpif19nIsjSdwpfA1meZfQICAPRAvZxCv+r9eewpuFuEw\nnm/FmYHXfq+dIxoi5MclNhEREZG9UdHNLq7WXk0eN+NiI8ZYIjYLnz2KIC3vsQ1wOAF7eFS78b1M\nnvldU6mzVL7AzQKq7QziUY4twGTq7fmk8w4OaoiQg88eh5/jYx5bREREpKMpKe7iwvSjwj5OOm/h\ntKX4mEKI/ds4isXh/zvGfPNwnjER3HYhKXxFkPhsdfFyDl57xs6kOBd9vERERKS7UNbSLaTSwA/a\nfbWhASJVUe0O00CqnR+3pLiRiwiFcRxPREREpOPFpSSbdG2WNDA5Ue0Rm0qQUQmIqCuypLCUFL4C\ngokORkRERDqZVoqTgoNI6nTwPobTeJtaQwzHz2EJjKtrcLKJXHNn475tgoToR529jACTEx2aiIiI\ndBIlxUnCpl9MnSedDN7GECBoh1PPT9AvCyDX3IvbrGp6ncJGcvgz5XYCKgcnIiKSHJQUJxE/J+C3\nJyQ6jC7FQTnO75S8A3CyhVTm4Oe4BEQlIiIinU3LhJLkDC2XpDOAs5NjERERkURRUixJLUIhIfpH\ntYfpq/3WIiIiSUTbJyQuDHVk82dcZj2WVBrsSfho72l2navG3kwed+BiIxAiTB9q7VWAO9GhiYiI\nSCdRUixxECLfXI/bLG9qcbEKh63Gy1kxjGt3/jf2E/f2JEJPKu3DONmMwUeIIeiXKCIiIslFSbHE\nLI33SWF1szan8ZDOu3jtmbQ1qXVQRq65HyebAScBewC1XENHr9yG6deh44uIiEjXpaRYYpbCMowJ\nRbUb6oAwbfuYhckzNzcrkda4guunxt4Wc6wiIiIiLdHviCVmfg4lYlOj2iMU0NbvXW4+x8X6Zm3G\nNJ42Z/C2fJGIiIhIjJQUS8wCHEKAsVj7zccpbAvw2LbvJ3ZShsNEH7Ns8GGojylOERERkd3R9glp\nM0MtDjyE6c2uOr/V9h4yeBU387Fk4LHnEGJYm8f2czhh+xxOU96sPUIPIvSIzw8gIiIi8h1KiqXV\nDA3kml/jYhUGPxF6UGcvI8BEwIWXM/DaM2K6R4QiGuw00nkNp6kBIGR7U2cvoaOrUIiIiEjyUlIs\nrZZjfkuamd302kkNOfyOCvs4lqy43aeei2iwx5FhXydCNl7+F0te3MYXERER+S4lxdJKIVJYEdXq\nMqWk2zdjrEccLcwg6rgqrmOKiIiI7I6SYmklS2N5tWgGz27aG0jnPzjZSgMnEGJEB8YnIiIi0n5K\nivcZlhQWkc47hOmJl1OxZMdx/BTC9MPF9matYVtAA9OjejtZT765DSebMMaSbt/FZ4+hluviGJOI\niIhIfKgk2z4ih9+Sb35FhuNNsh1PUWguw9XCdodY1NjrCdhhRGwKACHbC489kwg9o+MxD+EyGzGm\n8ahmh6knzbyPi6VxjUlEREQkHrRSvA9wsZQ08xEO8802BpfZQjZ/psr+IW73idCbSvsYbubhsJX4\nmYwlt8W+TrZFtTmMhwz7FrWMiltMIiIiIvGgpHgfkM57zRLiXZxsp3EvcDxLmTkIcOhee1miT7gD\nCNMrjrGIiIiIxIe2T+wDHOzA2uh2SxqJqu3rt4dgbfPvXCHbFy8/SEg8IiIiInuileJuzuAhxazC\nfCf3tRb89uDEBAXU81OwYVKZh8FHhF7U2ivjWs9YREREJF6UFHdzKSzcuU2iOYsTH8cnIKJdHNTz\nM+qtBSI4qCWT53CaZwjZgXg4TwmyiIiIdBlxSYoXLFjAU089RSQSYcqUKfzgB/oVeWeJUIQlHUND\ns3ZLLhHyExTVtxkc7KDAXI/LbNrZNIdU+zmV9sE4l40TERERaZ+Y9xRHIhGefPJJbrrpJn7/+98z\ne/ZsNm/eHI/YpBVC7EeQoVHtQYYToUcCIoqWbR7/JiHeKcWsIYtnEhSRiIiISHMxJ8WrV6+md+/e\n9OrVC5fLxeTJk5k3b148YpNWMVTbu/HZwwnZvoRsPxrsFGrs7YkOrImTshbbXWZDJ0ciIiIi0rKY\nt09UVlZSWFjY9LqwsJBVq1ZF9SspKaGkpASAGTNmUFRU1OZ7uVyudl237ysCHmNXCYoUYyj8To9E\nzp2jrgCC0e0pqX0oyuoef5767LWf5i42mr/209zFRvPXfpq72CRq/jrtQbupU6cyderUptfl5eVt\nHqOoqKhd10li587FWeSbpTjNN/cP2d5U+c4h7Osef5767LWf5i42mr/209zFRvPXfpq72MR7/oqL\ni1vVL+akuKCggIqKiqbXFRUVFBQUxDqs7ENCjKbK3kYWz+GglggF1NmLCdMn0aGJiIiIAHFIiocO\nHcrWrVspKyujoKCAOXPmcNVVV8UjNtmHhBhLtb0v0WGIiIiItCjmpNjpdHLRRRdx9913E4lEOPbY\nY+nfv388YhMRERER6RRx2VM8YcIEJkyYEI+hREREREQ6Xcwl2UREREREujslxSIiIiKS9JQUi4iI\niEjSU1IsIiIiIklPSbGIiIiIJD0lxSIiIiKS9JQUi4iIiEjSU1IsIiIiIklPSbGIiIiIJD0lxSIi\nIiKS9JQUi4iIiEjSU1IsIiIiIklPSbGIiIiIJD0lxSIiIiKS9JQUi4iIiEjSU1IsIiIiIklPSbGI\niIiIJD0lxSIiIiKS9JQUi4iIiEjSU1IsIiIiIklPSbGIiIiIJD0lxSIiIiKS9JQUi4iIiEjSU1Is\nIiIiIklPSbGIiIiIJD0lxSIiIiKS9JQUi4iIiEjSc8Vy8XPPPcf8+fNxuVz06tWLK664gszMzHjF\nJiIiIiLSKWJaKR47diwPPPAA999/P3369OHVV1+NV1wiIiIiIp0mpqR43LhxOJ1OAIYPH05lZWVc\nghIRERER6UzGWmvjMdCMGTOYPHkyRx11VIvvl5SUUFJS0tQ3EAi0+R4ul4tQKBRTnMlKcxcbzV/7\nae5io/lrP81dbPb1+WsIBnl64Vcs2lFGr4xMLjvoYHplZsVl7H197jpavOfP7Xa3qt9ek+K77rqL\n6urqqPazzz6bgw8+GIBXXnmFNWvWcN1112GMadWNS0tLW9Xv24qKiigvL2/zdaK5i5Xmr/00d7HR\n/LWf5i42+/L8+UIhrvpwFkurKpra+mZmMWPyUQzOzY15/H157jpDvOevuLi4Vf32+qDdrbfeusf3\nP/jgA+bPn89tt93W6oRYREREJFFeXLWiWUIMsMVTzyOLFvDbI45OUFSSaDHtKV6wYAGvvfYaN9xw\nA6mpqfGKSURERKTDLPtOQrxLua+hkyORriSmkmxPPvkkoVCIu+66C4Bhw4Zx6aWXxiUwERERkY7Q\nK6Pl8rFZKSmdHIl0JTElxX/84x/jFYeIiIhIp/jhiFF8urWUzZ76pra81FTOGjYigVFJosWUFIuI\niIh0NwVp6dx7+NE8smgBOxq8ZKakcPbwEUzu0zfRoUkCKSkWERGRpDMwJ4cZh7dcRlaSU0wP2omI\niIiI7AuUFIuIiIhI0lNSLCIiIiJJT0mxiIiIiCQ9JcUiIiIikvSUFIuIiIhI0lNSLCIiIiJJT0mx\niIiIiCQ9JcUiIiIikvR0op2IiIi0yjavh6eWLqLC52NwTi4XjhxNVoo70WGJxIWSYhEREdmrtTXV\n3DDnI7Z6PAB8um0rX2zfxp+OmUpmSkqCoxOJnbZPiIiIyF49uvjrpoR4l1U11bywYmmCIhKJLyXF\nIiIisleVPl+L7Wtqajo5EpGOoaRYRERE9ip7N3uHe2dkdHIkIh1DSbGIiIjs1fkjRlKYltasrX9W\nNheOHJ2giETiSw/aiYiIyF4d1LM3dxx6OM8tX0J9MEiP9AyuGDOegrT0RIcmEhdKikVERKRVxvfo\nyfgePRMdhkiHUFIsIiIikgALdpTx/IqleIMh+mZlccWY8eR/Z4sKwIdbNvHS6pV4g0F6ZmTys7Hj\n6ZuVnYCI921KikVERCTp+MMhIhbSXYlJhWaXbuHe+Z9T6W+s6rGwYgcrqip55Njjm9V9/u+mjfzu\nq3nUBAIArKiuYkNtDY8cezy5qakJiX1fpQftREREJGnUBgL8cvaHnPfOm5z3zptc/eEsyryevV8Y\nZ39fuawpId5lbW0N/1y1vFnby6tXNiXEu2ysr+NvK5d1eIzJRkmxiIiIJI1bP/2EOVtL2eb1Utbg\nZf6O7dw09xOstXG/VyAc5u0Na3l++VK2fSfxrvtOorvLuu/UffYEW+63pb4+PkFKE22fEBERkaSw\n3euJSjoB1tfWsLiynDGFPeJ2r/W1Ndwy9xM21tUSAf61egXfH7wfl4weA9Di1gcD7J9f0KytID2d\nNbXNY3YAY4uK4harNNJKsYiIiCQFTzCEPxyKaveHw9T6W16Rba/7vpzH+p0JMUCFz8dra1c1rfBe\nNGoMPdKbl7MbnpfP6fsNb9Z2+Zhx9MnIbHptgFGFRfxgyLC4xitaKRYREZEkMTAnm14Zmaz9zspr\ncWYWE3r2itt9AuEw2zzR+5Sr/H5mrl/DTw4Yx/gePbn/iGN4dtkSagMBhubm8uNRY0j7zoN/w/MK\nePiYqTyzbAkVvgYOKCzkjGH743Y64xavNFJSLCIiIknBaRxcPmYcv18wn9KdSWuvjAwuGDEqrlUo\nHMaQ4oz+ZbwB8lO/Kbk2NDePOyYdvtfxemZkcP1BB8ctPmlZXD4Bb7zxBs899xxPPPEEOTk58RhS\nREREJO4O69OXMUU9eHv9OgKRCCcNGtwsUY0Hl8PBAQVFbP7Ow3D9srKZNnhIXO8l8RNzUlxeXs7C\nhQsp0oZvERER6QayUtycPmz/Dr3HLw86hLC1LKkoJxCJ0Csjg6vHH0SGK2XvF0tCxJwUP/PMM5x3\n3nncd9998YhHREREpNtzO53cfuhk/OEwgXCYbLc70SHJXsSUFM+bN4+CggIGDRq0174lJSWUlJQA\nMGPGjHatLLtcLq1It5PmLjaav/bT3MVG89d+mrvYdOb8+UIhfKEQeS0ccdwd6bMXm0TN316T4rvu\nuovq6uqo9rPPPptXX32VW265pVU3mjp1KlOnTm16XV5e3oYwGxUVFbXrOtHcxUrz136au9ho/tpP\ncxebzpg/fzjMjC8+Y3FlOYFwmN4ZmVw9/iBGFRR26H07mj57sYn3/BUXF7eq316T4ltvvbXF9o0b\nN1JWVsb1118PQEVFBTfccAP33HMPeXl5bQhVREREktG98z/jvU0bml5X+Hzc+dkc/nr8idp7K52u\n3dsnBgwYwBNPPNH0+sorr+See+5R9QkRERHZq1AkwpKKiqj2zZ56HlrwJTdOPDQBUXU8ay1zt21l\nUfkOJvTsycSevTHGJDosQXWKRUREJAHCNkIwEmnxvVmbN3LJ6DEUpWd0clQdyx8Oce3HH7KssgJ/\nJMxLq1cyprCIe484ihTHvnkYRygSYWH5DtwOB6MKi3B04S8AcUuKH3744XgNJSIiIvu4VKeL4sxM\nyhq8Ue95QyH+uXIFV447MAGRdZy/LF7EgvKyptcN4RDzyrbxtxXLuHDkAQmMrGPM276VPyz4kk31\ndTiNYUB2DrcdchhDcrvmNtvo41ZEREREOsF1Ew7GtZuVQ08o2MnRdLzlVdHbRSywYEdZdOduLhAO\n8+CC+ayvqyVsLYFIhNU11dzzxWdYaxMdXou0fUJEREQSYlBOLof06sOcbaXN2nNS3Jy637BWj2Ot\nxRcOk+p0JuTX88FImL8sXsiiinIcxvC9YcP5fnH/qL3Cu9si4Xbue1snvijbxqa6+qj2LZ56Sj31\n9M3KTkBUe6akWERERBLmVxMP5ZezP2RtTQ3+SJii9HSmDRzMfrn5rbq+ZNMG/rZiGVV+H1kpKRzT\nbwAXjxrTwVE3d/PcT5i7tZRd65/LqipZPaiMXxw4sVm/U4YMZVlVBfXBb1bBc9xuztyvY0/XSwSn\nMTgMRL6zKGwwXXZfsZJiERERSZj8tDQeP+4E5pdtZ5vXw2F9iilMS2/VtWtqqvjj119S4fMBsKOh\nga0rlpHrdnN6JyWaa2uqWVRezrdzv0A4zKfbSvEGg2SkfFNa7th+A9ju9TJz/VrqggHy3Kn8z5D9\nOKhX706JtTMd1LM3/bNzWFdb06y9f1Y2fTKzEhTVnikpFhERkYQyxjCxHYnhCyuWNSXEu/jCYf67\naWPck2J/OAxYUp3NU6c1NdXUBQNR/WsDAXY0NDAwpXm95bOHj+DMYfvjC4dId7r22XJsLoeDGw86\nhPu+nEdpfT1Oh6H/zgftuiolxSIiItIteYOhFtt3V+qtPWr8fn4971PW1VZjLQzKyeHmgydRsHM1\ne0xhDwpS06j0N0/O81PT6J3Zckk5hzFJcTjJ6MIinpp6ImtqqnE7nQzI7tpnWaj6hIiISDfWEArR\nEGo5OdzXHVHcF5eJTmUGxTH5unnuJ8zdVso2r5ftDV4+276Nm+Z80lRBoXdmJpN698Ht+CaObLeb\nEwcOjlpVTkbGGPbLy+/yCTFopVhERKRbqvQ18OvPP2VDXS3QuIJ5y8GHkZ+WluDIOs+0QUOYs7WU\n+WXb8IRCpDgc7Jebz1XjD4rL+Jvr61hbWx3Vvq62mrW1NQzdWW/3VxMPZXyPnnyweRMOh+HC8Qcx\nMiOzXffcUFtLXTDA/vn5++yBHl2VkmIREZFu6KY5n7C4srzp9fYGL7d8+gkPHzM1gVF1Locx/Gby\nkSwq38GcraXsn5/PkX374Wxh9bg96gIBfC2swjeEQtT6/U2vjTFMGzSEaYOGAFBUVER5eXnUdXtS\n5fdx85xPWFdbTUMoTHFWJheNGsPU/gP3eN2ammrW19YwrqjHPncCYGdTUiwiItLNrK2piXqqf1f7\nxrrabvGr6ngaU9SDMUU94j7usLx8ijOzWL9zNX6XvlnZjCosiuu97p73KQsrdjS93lhXx6OLvubg\nnr3JTU2N6u8Ph7lpzscsraygLhigKC2No/r255rxB+2zD+91NO0pFhER6WbqgwF84egVTF84RF0g\nuhKC7Jk/HGJdbQ2eYPNT9FwOB5eMHkPvb22F6JWewY9GjiY1jgdu+MNh1rfwJWeb18Ob69e2eM3D\nC7/is+1bmypflPt8/Hvtan7y33f558rlBCPhuMWXLLRSLCIi0s2MyC+gX2YWG+rrmrX3zcxi//yC\nBEXVqLyhgTfXryHDlcK0QUPITOnaVRaeXb6Et9avo9LnI8ftZlLvPvziwIlNq63H9BvAfnl53PHZ\nXKp8PgrT05m3fRsWy3H9BibsNLplldFHRkesZXl1JSurK5mzrZTfHXlM3LaSJAMlxSIiIt2M2+nk\notFjeXTRArZ6PQD0ycjkktFjcDkSlwS9tGoFL6xYxg5fQ+Pr1Su5/qCDmdiz8w+nKK2v58EF89nq\n9ZDqdHJM3/6ct//IZlsL5m3byt9WLGs6Yc4TCvLm+rX0yczi3P1HAo2ryLfOnc2qmsYH7rY1eFlS\nWcE7G9fzworl/PqwwxmYnRtTrKlOJ4Nyctnm9TZr752Rwck79yl/155OhYsAC8t38MHmTUzZy55k\n+Ya+PoiIiHRDU/oP4MmpJ/LzsQdy9bgJ/HXqiRzTb0DC4qkNBPjHqhVNCTHAFk89D3/9VVP5ss7i\nC4X45ewPmbOtlHW1NSyvquTppYt5ZtmSZv1eWbuq2ZHLAIFIhNmlW5pev7pmFatroitQWGBdbQ33\nf/lFXGK++eBJjC0sIjslBZdxMDArm58eMK7F/cQAB/fsjWsPiXEwEmHu1tK4xJYstFIsIiLSTeW4\n3Zw1fESiwwBg7tYtbNu5av1t2xu8/7+9Ow+K8szzAP59++TqBrqbQzlEDo+RQxkR4gwmIMlumaRM\npTK7xuhqLDeOumWmyk1JdjSTrGG0xlInLrpa6piUbMXa2olmzcaRMlbGyTBJAC8OwQvxAGyaS7qB\nbvvYP9SOQANCIy9v+vv5r1/ffvvbP1vr108/7/OgwWJBVNDYbe37+fWr7qXqHulxOvD1nVtY/pNk\n99iLEzIAABC0SURBVDHHAM3648cvt7VhsJa+wWLGfadjwOXT7E4n/vtKLcqNd6GWy/GPSVORFhbe\n77xQtR/25DyPG/c6YL5/f8gl2VbMSIGppxtlxru422Xpl1GGBzcK0pNjU0xEREReC1X7QSWTwdZn\nNzmVTD7m84rrO+95bGS7+owK50bHouxuU6/MAoAUww8rS8wMC8dXt28O2EArBdmA83ZdLhfyS87g\n+6ZGPHqFi6Zm/DIlFS9NTvT4nDjt0FMxrrS3Yn/lRbRZrYgKDMJMQxj+0nAHFvsP72+yNgSvJDx4\nDcv9+/i09hLqOu9hakgIfpE0Df4KtoB9sSJERETktdkRkZikCcaVjrZex6eEhiJkgCkAT0tOdAyK\nb9b3W6EjLKD3Or5/FxuHc81G/K2pAa09PdAoVUjRG/BWcqr7nAVx8fiyvg6VLf3XHRYApBrCBpzf\ne7bZiAumZjz+NaHdZsUfr13Fi3EJD6Z1XKqC2WZDWEAA1qSkIXyITT9udd7DuyXf9BqVD/P3x+tT\npqLMeBc9Dgeig4KwLi0darkCbT09ePvMaVx/uLrFn+/cwtd3buOjebnQqFSDvpavYVNMREREA2qz\n9mBbeSnqO+9BKZMhPSwca1Jn9buh78FGGj/H1rLvcdvcCZkgYGqoDv82O3PMM88Oj8SciEiUNN6B\n/eEIb7i/P1Ylp/U6TxAEvDs7E01dFlSYTEgIDkb8w13qHlHIZPh9dg6KaqrxvbEJtzs7IQAIVCmR\nog/DOz/NGDBH2d0mj1twt1t7UGZswm9Lv/thDnYLcLmtFbtz8hCqHnhXwj9UV/abptLc3Y2a9jYU\neti45T8rzrsb4kcut7fhD9WVeHtm+oCv44vYFBMREZFHTpcL73zzZ9S0tbqP1XV0oMtuR76HZndC\nYBA+ejYXPXY7ZIIg2nJlLjxYNcL52JQHF4CB7kuLDAhEZOzAI7R+CgVWJqdiJR6MIHfabFDL5UO+\nvxl6PdQyGax9ppRolCp8WlvT66ZEALhp7sQnl6rwq0G2qW6z9ng83mn1vD71HYvZ4/E6D9tX+zqu\nPkFEREQefdvUgOt9Vl5wwIVzzUaP2x8/4qdQiNYQA8A3DbdxrtnYa9pCc3c39lVcGJXra1QqqORy\nuFwuNFrMMHV3eTxv7oQoTA3V9zoWoFDghdg496YbfTUO0MQ+EhOk8Xg8PMDzFs+BCs/zubWqsZ3S\nIgUcKSYiIiKPbnV29rtxDgC67Pdhvm+D3zi9WevMndsec5u6uz2cPTLXOtqxtew7NFgskAsCYjVa\nfJA1F3o/f/c5MkHA9uznsK/iPGrb26CSyfHi5Hi8EBuHC6Zmj9cdaovulTNSccHU3GtKRGyQBqtT\n0jyev3TaT1Db3oqWnh9GmMP8/bFs+ozhvF2fMD4/zURERCS6eVHR+K/aS2jt85N9mH9Ar+ZvvBlo\nBYfRWgXD4XLi378rwbXHGtNWaw9+8+1f+83r9Vco8KtZs/td463kFNR1tKPpsVHmeG0wlk4bvFkN\nVqtR+FweDtdU4WbnPUwMDMI/TZ8x4DzkFEMYfj07C0W11eiwWRGi9sOK6clI6DN3mtgUExER0QAm\nBAZhfkwsTtTXuTe5eDTKKAyycYTYXk1Mwp/q63DjsbWKAxQKPB8bNyrXP2u8i5t9ttgGHiwFd7fL\nAgMMHp7VW1KIDr+fl4uD1RVos/YgJkiDlTNSoX2CFSG0KhXWps564rxzIidgTuSEJz7fV7EpJiIi\nogG9PfOnyImOwf/WXUOQQonXp05HxBDLhj2JL29cx8n6G7A7nZiu0+Ofk1OhHqV5yAEKJbb9/Fl8\ndP4sGrvM8Jcr8ULsJLyaOGVUrm91OOHwMD3D7nTB5uh/fCDRGg1+kzl3VDKR99gUExER0aBSDeFI\nNfTfhW2kDlVX4tPLl9D18Ga9Cy3NuNrRhp3ZOaM2Aj0hMAhbfzZvVK7VV0ZEBKKCNLjVZ7R4YmAg\nosdw5z4aXWyKiYiIJMbhcuLj6iqUGpvgcrmQagjDquS0fmsHj0d2pxNf3ap3N8SPVLe24IKpGTM9\nbIE83qjlCqxOScOei+dxx2KGXBAQo9HiX9MzxvW0Ehocm2IiIiKJ+W3pd/jqVr17Y4pLrS1osJhR\n8Ey2yMmGds9mxT1b/+XIuux2VLeavGqK261W/F/ddchkwItxCU80P3ek5kXFYE7EBJQ03YFarkBm\nxIRx96XE9fDzwUb9yXjdFJ84cQInT56ETCZDeno6lixZMhq5iIiIyIO2nh6cbb7rbogBwAngoqkZ\njRYzJgSO75/vg9VqhKjV/Va0CFIqMSssYsTX/VP9deyvrMDdh6s5fHb1Cv4lbRaejYrxKu9g/BQK\n5EZPemrXHymrw47flZfiUmsLnC4XJmm1yJ+dOehOeeRlU1xZWYmysjJs27YNSqUSHR0dQz+JiIiI\nRqypy4IOq7Xf8TarFbfNneO+KZYLMrw0OR6HqqvcG1jIIWCmIRzTdfohnu2Z1WHHJ5eq3Q0xADR2\nWbC/6iJ+NiFq3I3gPm3vf/c3/KXhtvvxbYsZG/56Bvtynueo8SC8aoqLi4uxcOFCKB+u+xcc7Hld\nQCIiIhodkzRaGPwD0NBn57Nw/wAkhoSKlGp4/iFpGuI0wfjjtSu473QgIyISv0icOuLrVbWYcMfD\nEmmNZguudrRjWqjOm7iS0mbtQU1rS7/j1zs6UNFiQqohTIRU0uBVU9zY2IiamhocOXIESqUSS5cu\nRWJiosdzT506hVOnTgEAtm7dCoNh6DX8+oVVKEb0PGLtvMX6jRxr5x3Wb+R+zLVbOHU6iiouuEda\nAxQK/H1CEpKiokftNZ52/RYYDFiQnDIq14qFC/5KJSwP11J+xF+pxKSICBgG2MzjaRHzs3evrRU9\nHpaL63HY0aOQS+LfhFj1G7Ip3rx5M9rb2/sdX7RoEZxOJ8xmMwoKCnDt2jXs3LkThYWFHofm8/Ly\nkJf3wy4vJpNp2GENBsOInkesnbdYv5Fj7bzD+o3cj7l2b8QnIjEgEJ9fvwonXFgwaTLmRcWM6vuV\nUv30LiBOo0VVnxHSyVot/G33x/x9iFm7QJcTYX5+6LT1nmITGRCAaQFBkvg7He36TZw48YnOG7Ip\n3rRp04B/VlxcjDlz5kAQBCQmJkImk6GzsxNa7eD7dhMREZF3MiMnIJO7lAF4sLrCb+dmo6D0W9Tf\nuwdBEBCvDcavM7LEjjbm5IIMb05Pxn9cPAfjwznWoWo1XolPeqqrcfwYeDV9IiMjA1VVVUhOTkZD\nQwPsdjs0Gs1oZSMiIiJ6Ino/f+zIzoHVYQcgjNrueFKUExOLFEMY/ufqZVgddryakIQYDQcsh+JV\nU5ybm4s9e/Zg/fr1UCgUWLt2Le9qJCIiItGo5dyCAQAM/v74ZUqa2DEkxatPjkKhwLp160YrCxER\nERGRKHxr4T4iIiIiIg/YFBMRERGRz2NTTEREREQ+j00xEREREfk8NsVERERE5PPYFBMRERGRz2NT\nTEREREQ+T3C5XC6xQxARERERiUlSI8X5+fliR5As1s47rN/IsXbeYf1GjrXzDus3cqydd8Sqn6Sa\nYiIiIiKip4FNMRERERH5PPn777//vtghhiM+Pl7sCJLF2nmH9Rs51s47rN/IsXbeYf1GjrXzjhj1\n4412REREROTzOH2CiIiIiHwem2IiIiIi8nkKsQMMx40bN7B//37YbDbI5XKsXLkSiYmJYseSlBMn\nTuDkyZOQyWRIT0/HkiVLxI4kKcePH8fhw4dx4MABaLVaseNIxuHDh1FeXg6FQoGIiAisWbMGgYGB\nYsca186fP49Dhw7B6XRi/vz5eOWVV8SOJBkmkwm7d+9Ge3s7BEFAXl4eFixYIHYsSXE6ncjPz4dO\np+PyYsNksViwd+9e3Lp1C4IgYPXq1ZgyZYrYsSThiy++wOnTpyEIAmJiYrBmzRqoVKoxe31JNcVF\nRUV47bXXMGvWLJw9exZFRUWQ2H2CoqqsrERZWRm2bdsGpVKJjo4OsSNJislkwsWLF2EwGMSOIjmp\nqalYvHgx5HI5ioqKcPToUX4hG4TT6cTBgwexceNG6PV6vPvuu5g9ezaio6PFjiYJcrkcS5cuRXx8\nPLq7u5Gfn4/U1FTWbxi+/PJLREVFobu7W+woknPo0CHMnDkT69evh91uh9VqFTuSJLS2tuLEiRPY\nuXMnVCoVduzYgZKSEjz33HNjlkFS0ycEQXD/A+3q6kJoaKjIiaSluLgYCxcuhFKpBAAEBweLnEha\nPvnkE7zxxhsQBEHsKJKTlpYGuVwOAJgyZQpaW1tFTjS+Xb16FZGRkYiIiIBCocDcuXNRWloqdizJ\nCA0Ndd+57u/vj6ioKH7mhqGlpQVnz57F/PnzxY4iOV1dXbh06RJyc3MBAAqFgr+KDYPT6YTNZoPD\n4YDNZhvzPk9SI8XLli1DQUEBDh8+DKfTiQ8//FDsSJLS2NiImpoaHDlyBEqlEkuXLuX0kydUWloK\nnU6HuLg4saNI3unTpzF37lyxY4xrra2t0Ov17sd6vR5XrlwRMZF0GY1G1NXV8f+6Yfj444+xZMkS\njhKPgNFohFarxZ49e1BfX4/4+HgsX74cfn5+Ykcb93Q6HV5++WWsXr0aKpUKaWlpSEtLG9MM464p\n3rx5M9rb2/sdX7RoESoqKrBs2TJkZWWhpKQEe/fuxaZNm0RIOX4NVj+n0wmz2YyCggJcu3YNO3fu\nRGFhIUc+HxqsdkePHsXGjRtFSCUdg9UvIyMDAPDZZ59BLpcjOzt7rOORD+rp6cH27duxfPlyBAQE\niB1HEsrLyxEcHIz4+HhUVVWJHUdyHA4H6urqsGLFCiQlJeHQoUM4duwYFi1aJHa0cc9sNqO0tBS7\nd+9GQEAAduzYgTNnzmDevHljlmHcNcWDNbmFhYV48803AQDPPPMM9u3bN1axJGOw+hUXF2POnDkQ\nBAGJiYmQyWTo7OzkDWMPDVS7mzdvwmg04p133gHw4KfFDRs2YMuWLQgJCRnLiOPaUF9Qv/76a5SX\nl+O9997jF7Eh6HQ6tLS0uB+3tLRAp9OJmEh67HY7tm/fjuzsbGRmZoodRzJqa2tRVlaGc+fOwWaz\nobu7G7t27cK6devEjiYJer0eer0eSUlJAICsrCwcO3ZM5FTSUFFRgfDwcHdPkpmZicuXL/t2UzwY\nnU6H6upqzJgxA5WVlYiMjBQ7kqRkZGSgqqoKycnJaGhogN1uh0ajETvWuBcbG4sDBw64H69duxZb\ntmzhl4lhOH/+PD7//HN88MEHUKvVYscZ9xISEtDY2Aij0QidToeSkhI2JcPgcrmwd+9eREVF4aWX\nXhI7jqQsXrwYixcvBgBUVVXh+PHj/OwNQ0hICPR6PRoaGjBx4kRUVFTwBs8nZDAYcOXKFVitVqhU\nKlRUVCAhIWFMM0iqKV61apV7iSKlUolVq1aJHUlScnNzsWfPHqxfvx4KhQJr167liB2NiYMHD8Ju\nt2Pz5s0AgKSkJLz11lsipxq/5HI5VqxYgYKCAjidTuTk5CAmJkbsWJJRW1uLM2fOIDY21v0Lz+uv\nv4709HSRk5EvWLFiBXbt2gW73Y7w8HCsWbNG7EiSkJSUhKysLGzYsAFyuRxxcXHIy8sb0wzc5pmI\niIiIfJ6klmQjIiIiInoa2BQTERERkc9jU0xEREREPo9NMRERERH5PDbFREREROTz2BQTERERkc9j\nU0xEREREPu//AVb/wPKBs0gYAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xcea8470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Работоспособность KMeans\n",
    "km = Kmeans(3, random_state=Rnd_state)\n",
    "km.fit(X)\n",
    "plt.scatter(X[:,0], X[:, 1], c=km.labels)\n",
    "print km.centroids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-5.99382611  2.25004374]\n",
      " [-1.29992907  5.919722  ]\n",
      " [ 5.81037396 -4.9431358 ]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsUAAAEyCAYAAAAStK0kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIhIzFNTLCIiIiIx7/8Bk0tZKq1JjzAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xcdf7208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Работоспособность KMeans++\n",
    "km = Kmeans(3, random_state=Rnd_state, init='k-means')\n",
    "km.fit(X)\n",
    "plt.scatter(X[:,0], X[:, 1], c=km.labels)\n",
    "print km.centroids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-1.25273357  5.7322764 ]\n",
      " [ 5.63574659 -4.78340524]\n",
      " [-5.81873958  2.19199256]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsUAAAEyCAYAAAAStK0kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nU0xEREREPu//AVb/wPKBs0gYAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xce9de48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Работоспособность KMeans с MiniBatchKMeans\n",
    "km = MiniBatchKMeans(3, random_state=Rnd_state, batch_size=50)\n",
    "km.fit(X)\n",
    "labels = km.predict(X)\n",
    "plt.scatter(X[:,0], X[:, 1], c=labels)\n",
    "print km.centroids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[-5.81788307  2.193091  ]\n",
      " [-1.25028528  5.73066411]\n",
      " [ 5.62954395 -4.78607547]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsUAAAEyCAYAAAAStK0kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niIhIzFNTLCIiIiIx7/8Bk0tZKq1JjzAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xe0e69e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## Работоспособность KMeans++ с MiniBatchKMeans\n",
    "km = MiniBatchKMeans(3, random_state=Rnd_state, batch_size=50, init=\"k-means\")\n",
    "km.fit(X)\n",
    "labels = km.predict(X)\n",
    "plt.scatter(X[:,0], X[:, 1], c=labels)\n",
    "print km.centroids"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Время выполнения KMeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1250ba90>]"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsYAAAEyCAYAAAD5gxYnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJEmSBFiMJUmSJAD+f60sc8IFlcwNAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x122c86a0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "resy = []\n",
    "resx = []\n",
    "for i in xrange(1, 5):\n",
    "    X, y = make_blobs(n_samples= 5**i * 10, n_features=2, centers=3, cluster_std=1, \n",
    "           center_box=(-10.0, 10.0), shuffle=False, random_state=1234)\n",
    "    km = Kmeans(3, random_state=Rnd_state, max_iter=100)\n",
    "    tt = %timeit -n 1 -q -o km.fit(X)\n",
    "    \n",
    "    resy.append(tt.best)\n",
    "    resx.append(5**i * 10)\n",
    "\n",
    "plt.plot(resx, resy)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Время выполнения алгоритма KMeans++"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x12723780>]"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsYAAAEzCAYAAAAy38WCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x123bb978>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "resy = []\n",
    "resx = []\n",
    "for i in xrange(1, 5):\n",
    "    X, y = make_blobs(n_samples= 5**i * 10, n_features=2, centers=3, cluster_std=1, \n",
    "           center_box=(-10.0, 10.0), shuffle=False, random_state=1234)\n",
    "    km = Kmeans(3, random_state=Rnd_state, max_iter=100, init='k-means')\n",
    "    tt = %timeit -n 1 -q -o km.fit(X)\n",
    "    \n",
    "    resy.append(tt.best)\n",
    "    resx.append(5**i * 10)\n",
    "\n",
    "plt.plot(resx, resy)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Время выполнения алгоритма KMeans с MiniBatchKMeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x12b13a20>]"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAswAAAEyCAYAAADuoYbuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYRYRERERuYH/HxukPIZWrCs5AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x12a2c208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "resy = []\n",
    "resx = []\n",
    "for i in xrange(2, 6):\n",
    "    X, y = make_blobs(n_samples= 5**i * 10, n_features=2, centers=3, cluster_std=1, \n",
    "           center_box=(-10.0, 10.0), shuffle=False, random_state=1234)\n",
    "    km = MiniBatchKMeans(3, random_state=Rnd_state, batch_size=5**i, max_iter=100)\n",
    "    tt = %timeit -n 1 -q -o km.fit(X)\n",
    "    \n",
    "    resy.append(tt.best)\n",
    "    resx.append(5**i * 10)\n",
    "\n",
    "plt.plot(resx, resy)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Время выполнения алгоритма  KMeans++ с MiniBatchKMeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x130f7438>]"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAswAAAEyCAYAAADuoYbuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xe49f438>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "resy = []\n",
    "resx = []\n",
    "for i in xrange(2, 6):\n",
    "    X, y = make_blobs(n_samples= 5**i * 10, n_features=2, centers=3, cluster_std=1, \n",
    "           center_box=(-10.0, 10.0), shuffle=False, random_state=1234)\n",
    "    km = MiniBatchKMeans(3, random_state=Rnd_state, batch_size=5**i, max_iter=100, init='k-means')\n",
    "    tt = %timeit -n 1 -q -o km.fit(X)\n",
    "    \n",
    "    resy.append(tt.best)\n",
    "    resx.append(5**i * 10)\n",
    "\n",
    "plt.plot(resx, resy)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Применение K-means на реальных данных"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Загрузите [данные](https://github.com/brenden17/sklearnlab/blob/master/facebook/snsdata.csv) в которых содержится описание интересов профилей учеников старшей школы США.  (без этого пункта задание не считается выполненным).   \n",
    "ВНИМАНИЕ! Проверка должна быть осуществлена на всех реализациях, иначе реализация не будет зачтена!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "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>basketball</th>\n",
       "      <th>football</th>\n",
       "      <th>soccer</th>\n",
       "      <th>softball</th>\n",
       "      <th>volleyball</th>\n",
       "      <th>swimming</th>\n",
       "      <th>cheerleading</th>\n",
       "      <th>baseball</th>\n",
       "      <th>tennis</th>\n",
       "      <th>sports</th>\n",
       "      <th>...</th>\n",
       "      <th>blonde</th>\n",
       "      <th>mall</th>\n",
       "      <th>shopping</th>\n",
       "      <th>clothes</th>\n",
       "      <th>hollister</th>\n",
       "      <th>abercrombie</th>\n",
       "      <th>die</th>\n",
       "      <th>death</th>\n",
       "      <th>drunk</th>\n",
       "      <th>drugs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 36 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   basketball  football  soccer  softball  volleyball  swimming  cheerleading  \\\n",
       "0           0         0       0         0           0         0             0   \n",
       "1           0         1       0         0           0         0             0   \n",
       "2           0         1       0         0           0         0             0   \n",
       "3           0         0       0         0           0         0             0   \n",
       "4           0         0       0         0           0         0             0   \n",
       "\n",
       "   baseball  tennis  sports  ...    blonde  mall  shopping  clothes  \\\n",
       "0         0       0       0  ...         0     0         0        0   \n",
       "1         0       0       0  ...         0     1         0        0   \n",
       "2         0       0       0  ...         0     0         0        0   \n",
       "3         0       0       0  ...         0     0         0        0   \n",
       "4         0       0       0  ...         0     0         2        0   \n",
       "\n",
       "   hollister  abercrombie  die  death  drunk  drugs  \n",
       "0          0            0    0      0      0      0  \n",
       "1          0            0    0      0      0      0  \n",
       "2          0            0    0      1      0      0  \n",
       "3          0            0    0      0      0      0  \n",
       "4          0            0    0      0      1      1  \n",
       "\n",
       "[5 rows x 36 columns]"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sns = pd.read_csv('snsdata.csv', sep=',')\n",
    "df_sns.shape\n",
    "df_sns_users = df_sns.iloc[:, 0:4]\n",
    "df_sns_words = df_sns.iloc[:, 4:]\n",
    "df_sns_words1 = (df_sns_words - df_sns_words.mean(axis=0))/(df_sns_words.std(axis=0))\n",
    "df_sns_words.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Данные устроены так: \n",
    "* Год выпуска\n",
    "* Пол\n",
    "* Возраст\n",
    "* Количество друзей\n",
    "* 36 ключевых слов, которые встречаются в профилe facebook (интересы, сообщества, встречи)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Удалите все признаки кроме 36 ключевых слов.\n",
    "* Нормализуйте данные - из каждого столбца вычтите его среднее значение и поделите на стандартное отклонение.\n",
    "* Используйте метод k-means чтобы выделить 9 кластеров\n",
    "* Попробуйте проинтерпретировать каждый кластер проанализировав полученные центройды (Некоторые кластеры могут быть очень большие и очень маленькие - плохо интерпретируются)   \n",
    "КОММЕНТАРИИ НЕ СТИРАТЬ!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Kmeans(init='random', k=9, max_iter=100, metric='euclidean',\n",
       "    random_state=None)"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## KMeans\n",
    "km = Kmeans(9, max_iter=100)\n",
    "km.fit(df_sns_words1.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Centroid 1\n",
      "Element count 1418\n",
      "basketball      1.880255\n",
      "football        0.649745\n",
      "soccer          0.471896\n",
      "softball        1.762056\n",
      "volleyball      1.248601\n",
      "swimming        0.156724\n",
      "cheerleading    0.140568\n",
      "baseball        0.171222\n",
      "tennis          0.132974\n",
      "sports          0.698969\n",
      "cute            0.308755\n",
      "sex             0.151340\n",
      "sexy            0.097142\n",
      "hot             0.150165\n",
      "kissed          0.077741\n",
      "dance           0.486881\n",
      "band            0.241164\n",
      "marching        0.026857\n",
      "music           0.761184\n",
      "rock            0.247376\n",
      "god             0.445872\n",
      "church          0.288663\n",
      "jesus           0.094864\n",
      "bible           0.011047\n",
      "hair            0.472313\n",
      "dress           0.110812\n",
      "blonde          0.122894\n",
      "mall            0.275476\n",
      "shopping        0.427643\n",
      "clothes         0.192902\n",
      "hollister       0.039078\n",
      "abercrombie     0.028446\n",
      "die             0.143262\n",
      "death           0.121099\n",
      "drunk           0.063312\n",
      "drugs           0.036592\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 2\n",
      "Element count 2389\n",
      "basketball      0.228717\n",
      "football        0.878007\n",
      "soccer          0.276386\n",
      "softball        0.088004\n",
      "volleyball      0.103497\n",
      "swimming        0.133049\n",
      "cheerleading    0.101898\n",
      "baseball        0.076116\n",
      "tennis          0.233715\n",
      "sports          0.145878\n",
      "cute            0.285036\n",
      "sex             0.132313\n",
      "sexy            0.057345\n",
      "hot             0.083822\n",
      "kissed          0.055008\n",
      "dance           0.468868\n",
      "band            0.469276\n",
      "marching        0.068410\n",
      "music           0.887689\n",
      "rock            0.259335\n",
      "god             1.656536\n",
      "church          1.027122\n",
      "jesus           0.605271\n",
      "bible           0.133006\n",
      "hair            0.383367\n",
      "dress           0.104029\n",
      "blonde          0.084109\n",
      "mall            0.205969\n",
      "shopping        0.235031\n",
      "clothes         0.113252\n",
      "hollister       0.023282\n",
      "abercrombie     0.012584\n",
      "die             0.227363\n",
      "death           0.176742\n",
      "drunk           0.061074\n",
      "drugs           0.033488\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 3\n",
      "Element count 18385\n",
      "basketball      0.120851\n",
      "football        0.098591\n",
      "soccer          0.136345\n",
      "softball        0.056023\n",
      "volleyball      0.058009\n",
      "swimming        0.066911\n",
      "cheerleading    0.046480\n",
      "baseball        0.030571\n",
      "tennis          0.052416\n",
      "sports          0.072352\n",
      "cute            0.165495\n",
      "sex             0.094995\n",
      "sexy            0.064903\n",
      "hot             0.060549\n",
      "kissed          0.034026\n",
      "dance           0.239719\n",
      "band            0.163396\n",
      "marching        0.015689\n",
      "music           0.513844\n",
      "rock            0.135602\n",
      "god             0.232045\n",
      "church          0.070858\n",
      "jesus           0.033036\n",
      "bible           0.000883\n",
      "hair            0.189344\n",
      "dress           0.040383\n",
      "blonde          0.040490\n",
      "mall            0.117595\n",
      "shopping        0.126143\n",
      "clothes         0.053284\n",
      "hollister       0.014199\n",
      "abercrombie     0.009252\n",
      "die             0.119606\n",
      "death           0.073466\n",
      "drunk           0.053676\n",
      "drugs           0.021281\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 4\n",
      "Element count 311\n",
      "basketball      0.392043\n",
      "football        0.313717\n",
      "soccer          0.239448\n",
      "softball        0.244723\n",
      "volleyball      0.133117\n",
      "swimming        0.183441\n",
      "cheerleading    0.237417\n",
      "baseball        0.112419\n",
      "tennis          0.124594\n",
      "sports          0.189123\n",
      "cute            0.440745\n",
      "sex             0.185065\n",
      "sexy            0.134335\n",
      "hot             0.189934\n",
      "kissed          0.127840\n",
      "dance           0.538147\n",
      "band            0.472400\n",
      "marching        0.045049\n",
      "music           1.058030\n",
      "rock            0.323456\n",
      "god             4.803093\n",
      "church          4.973136\n",
      "jesus           1.900538\n",
      "bible           0.732943\n",
      "hair            0.538147\n",
      "dress           0.195615\n",
      "blonde          0.108766\n",
      "mall            0.424104\n",
      "shopping        0.665985\n",
      "clothes         0.327107\n",
      "hollister       0.063718\n",
      "abercrombie     0.019481\n",
      "die             0.387578\n",
      "death           0.245939\n",
      "drunk           0.109983\n",
      "drugs           0.076298\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 5\n",
      "Element count 3592\n",
      "basketball      0.236058\n",
      "football        0.255760\n",
      "soccer          0.441478\n",
      "softball        0.099291\n",
      "volleyball      0.141887\n",
      "swimming        0.413922\n",
      "cheerleading    0.318229\n",
      "baseball        0.061525\n",
      "tennis          0.120142\n",
      "sports          0.118446\n",
      "cute            0.741213\n",
      "sex             0.142406\n",
      "sexy            0.094746\n",
      "hot             0.241635\n",
      "kissed          0.071102\n",
      "dance           0.812517\n",
      "band            0.230881\n",
      "marching        0.017488\n",
      "music           0.858772\n",
      "rock            0.220353\n",
      "god             0.404753\n",
      "church          0.256939\n",
      "jesus           0.066069\n",
      "bible           0.004487\n",
      "hair            0.620815\n",
      "dress           0.400691\n",
      "blonde          0.114677\n",
      "mall            0.775669\n",
      "shopping        1.438070\n",
      "clothes         0.408343\n",
      "hollister       0.069347\n",
      "abercrombie     0.035091\n",
      "die             0.166422\n",
      "death           0.123910\n",
      "drunk           0.082089\n",
      "drugs           0.032963\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 6\n",
      "Element count 493\n",
      "basketball      0.816874\n",
      "football        0.939873\n",
      "soccer          0.367084\n",
      "softball        0.220208\n",
      "volleyball      0.123641\n",
      "swimming        0.179279\n",
      "cheerleading    0.208909\n",
      "baseball        3.201824\n",
      "tennis          0.094649\n",
      "sports          0.437002\n",
      "cute            0.399913\n",
      "sex             0.280323\n",
      "sexy            0.146664\n",
      "hot             0.163078\n",
      "kissed          0.095076\n",
      "dance           0.412705\n",
      "band            0.209978\n",
      "marching        0.015988\n",
      "music           0.811979\n",
      "rock            0.593684\n",
      "god             0.445534\n",
      "church          0.364739\n",
      "jesus           0.091878\n",
      "bible           0.015349\n",
      "hair            0.510764\n",
      "dress           0.043062\n",
      "blonde          0.330203\n",
      "mall            0.289916\n",
      "shopping        0.201452\n",
      "clothes         0.129397\n",
      "hollister       0.029206\n",
      "abercrombie     0.022597\n",
      "die             0.146878\n",
      "death           0.106374\n",
      "drunk           0.054573\n",
      "drugs           0.053294\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 7\n",
      "Element count 889\n",
      "basketball      0.500166\n",
      "football        0.539076\n",
      "soccer          0.323992\n",
      "softball        0.238229\n",
      "volleyball      0.195915\n",
      "swimming        0.278387\n",
      "cheerleading    0.250026\n",
      "baseball        0.188201\n",
      "tennis          0.156777\n",
      "sports          0.493811\n",
      "cute            0.713890\n",
      "sex             2.600641\n",
      "sexy            0.324898\n",
      "hot             0.270673\n",
      "kissed          1.790670\n",
      "dance           0.930905\n",
      "band            0.757565\n",
      "marching        0.056721\n",
      "music           2.041725\n",
      "rock            0.929087\n",
      "god             0.924099\n",
      "church          0.364037\n",
      "jesus           0.143959\n",
      "bible           0.034373\n",
      "hair            3.283219\n",
      "dress           0.373904\n",
      "blonde          0.888019\n",
      "mall            0.701184\n",
      "shopping        0.544523\n",
      "clothes         0.737937\n",
      "hollister       0.133181\n",
      "abercrombie     0.135563\n",
      "die             1.305597\n",
      "death           0.534536\n",
      "drunk           0.937365\n",
      "drugs           1.094935\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 8\n",
      "Element count 859\n",
      "basketball      0.385742\n",
      "football        0.407691\n",
      "soccer          0.322434\n",
      "softball        0.164279\n",
      "volleyball      0.256583\n",
      "swimming        0.244684\n",
      "cheerleading    0.313190\n",
      "baseball        0.103397\n",
      "tennis          0.150415\n",
      "sports          0.179412\n",
      "cute            0.620374\n",
      "sex             0.240411\n",
      "sexy            0.192351\n",
      "hot             0.334331\n",
      "kissed          0.144292\n",
      "dance           0.687958\n",
      "band            0.205755\n",
      "marching        0.014903\n",
      "music           0.905612\n",
      "rock            0.296210\n",
      "god             0.467999\n",
      "church          0.231632\n",
      "jesus           0.113332\n",
      "bible           0.015827\n",
      "hair            0.939917\n",
      "dress           0.173636\n",
      "blonde          0.214416\n",
      "mall            0.657456\n",
      "shopping        0.904566\n",
      "clothes         0.404917\n",
      "hollister       1.493391\n",
      "abercrombie     1.158369\n",
      "die             0.219731\n",
      "death           0.158618\n",
      "drunk           0.097158\n",
      "drugs           0.078789\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 9\n",
      "Element count 1664\n",
      "basketball      0.268228\n",
      "football        0.277146\n",
      "soccer          0.267411\n",
      "softball        0.116750\n",
      "volleyball      0.097533\n",
      "swimming        0.147836\n",
      "cheerleading    0.102242\n",
      "baseball        0.087170\n",
      "tennis          0.086479\n",
      "sports          0.138919\n",
      "cute            0.856048\n",
      "sex             0.469882\n",
      "sexy            1.135448\n",
      "hot             0.453489\n",
      "kissed          0.131005\n",
      "dance           1.185631\n",
      "band            1.650232\n",
      "marching        0.371658\n",
      "music           1.934477\n",
      "rock            0.996094\n",
      "god             0.544745\n",
      "church          0.221692\n",
      "jesus           0.087296\n",
      "bible           0.011556\n",
      "hair            0.766433\n",
      "dress           0.138604\n",
      "blonde          0.207685\n",
      "mall            0.293411\n",
      "shopping        0.294167\n",
      "clothes         0.235758\n",
      "hollister       0.031402\n",
      "abercrombie     0.017334\n",
      "die             0.288575\n",
      "death           0.195251\n",
      "drunk           0.122966\n",
      "drugs           0.057590\n",
      "dtype: float64\n",
      "==========================\n"
     ]
    }
   ],
   "source": [
    "for i in xrange(9):\n",
    "    print \"Centroid \" + str(i+1)\n",
    "    print \"Element count \" + str(sum([1 for  j in km.labels if j == i]))\n",
    "    print km.centroids[i] * df_sns_words.values.std(axis=0) + df_sns_words.mean(axis=0)\n",
    "    print \"==========================\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Kmeans(init='k-means', k=9, max_iter=100, metric='euclidean',\n",
       "    random_state=None)"
      ]
     },
     "execution_count": 152,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## KMeans++\n",
    "km = Kmeans(9, max_iter=100, init=\"k-means\")\n",
    "km.fit(df_sns_words1.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Centroid 1\n",
      "Element count 1418\n",
      "basketball      1.880255\n",
      "football        0.649745\n",
      "soccer          0.471896\n",
      "softball        1.762056\n",
      "volleyball      1.248601\n",
      "swimming        0.156724\n",
      "cheerleading    0.140568\n",
      "baseball        0.171222\n",
      "tennis          0.132974\n",
      "sports          0.698969\n",
      "cute            0.308755\n",
      "sex             0.151340\n",
      "sexy            0.097142\n",
      "hot             0.150165\n",
      "kissed          0.077741\n",
      "dance           0.486881\n",
      "band            0.241164\n",
      "marching        0.026857\n",
      "music           0.761184\n",
      "rock            0.247376\n",
      "god             0.445872\n",
      "church          0.288663\n",
      "jesus           0.094864\n",
      "bible           0.011047\n",
      "hair            0.472313\n",
      "dress           0.110812\n",
      "blonde          0.122894\n",
      "mall            0.275476\n",
      "shopping        0.427643\n",
      "clothes         0.192902\n",
      "hollister       0.039078\n",
      "abercrombie     0.028446\n",
      "die             0.143262\n",
      "death           0.121099\n",
      "drunk           0.063312\n",
      "drugs           0.036592\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 2\n",
      "Element count 2389\n",
      "basketball      0.228717\n",
      "football        0.878007\n",
      "soccer          0.276386\n",
      "softball        0.088004\n",
      "volleyball      0.103497\n",
      "swimming        0.133049\n",
      "cheerleading    0.101898\n",
      "baseball        0.076116\n",
      "tennis          0.233715\n",
      "sports          0.145878\n",
      "cute            0.285036\n",
      "sex             0.132313\n",
      "sexy            0.057345\n",
      "hot             0.083822\n",
      "kissed          0.055008\n",
      "dance           0.468868\n",
      "band            0.469276\n",
      "marching        0.068410\n",
      "music           0.887689\n",
      "rock            0.259335\n",
      "god             1.656536\n",
      "church          1.027122\n",
      "jesus           0.605271\n",
      "bible           0.133006\n",
      "hair            0.383367\n",
      "dress           0.104029\n",
      "blonde          0.084109\n",
      "mall            0.205969\n",
      "shopping        0.235031\n",
      "clothes         0.113252\n",
      "hollister       0.023282\n",
      "abercrombie     0.012584\n",
      "die             0.227363\n",
      "death           0.176742\n",
      "drunk           0.061074\n",
      "drugs           0.033488\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 3\n",
      "Element count 18385\n",
      "basketball      0.120851\n",
      "football        0.098591\n",
      "soccer          0.136345\n",
      "softball        0.056023\n",
      "volleyball      0.058009\n",
      "swimming        0.066911\n",
      "cheerleading    0.046480\n",
      "baseball        0.030571\n",
      "tennis          0.052416\n",
      "sports          0.072352\n",
      "cute            0.165495\n",
      "sex             0.094995\n",
      "sexy            0.064903\n",
      "hot             0.060549\n",
      "kissed          0.034026\n",
      "dance           0.239719\n",
      "band            0.163396\n",
      "marching        0.015689\n",
      "music           0.513844\n",
      "rock            0.135602\n",
      "god             0.232045\n",
      "church          0.070858\n",
      "jesus           0.033036\n",
      "bible           0.000883\n",
      "hair            0.189344\n",
      "dress           0.040383\n",
      "blonde          0.040490\n",
      "mall            0.117595\n",
      "shopping        0.126143\n",
      "clothes         0.053284\n",
      "hollister       0.014199\n",
      "abercrombie     0.009252\n",
      "die             0.119606\n",
      "death           0.073466\n",
      "drunk           0.053676\n",
      "drugs           0.021281\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 4\n",
      "Element count 311\n",
      "basketball      0.392043\n",
      "football        0.313717\n",
      "soccer          0.239448\n",
      "softball        0.244723\n",
      "volleyball      0.133117\n",
      "swimming        0.183441\n",
      "cheerleading    0.237417\n",
      "baseball        0.112419\n",
      "tennis          0.124594\n",
      "sports          0.189123\n",
      "cute            0.440745\n",
      "sex             0.185065\n",
      "sexy            0.134335\n",
      "hot             0.189934\n",
      "kissed          0.127840\n",
      "dance           0.538147\n",
      "band            0.472400\n",
      "marching        0.045049\n",
      "music           1.058030\n",
      "rock            0.323456\n",
      "god             4.803093\n",
      "church          4.973136\n",
      "jesus           1.900538\n",
      "bible           0.732943\n",
      "hair            0.538147\n",
      "dress           0.195615\n",
      "blonde          0.108766\n",
      "mall            0.424104\n",
      "shopping        0.665985\n",
      "clothes         0.327107\n",
      "hollister       0.063718\n",
      "abercrombie     0.019481\n",
      "die             0.387578\n",
      "death           0.245939\n",
      "drunk           0.109983\n",
      "drugs           0.076298\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 5\n",
      "Element count 3592\n",
      "basketball      0.236058\n",
      "football        0.255760\n",
      "soccer          0.441478\n",
      "softball        0.099291\n",
      "volleyball      0.141887\n",
      "swimming        0.413922\n",
      "cheerleading    0.318229\n",
      "baseball        0.061525\n",
      "tennis          0.120142\n",
      "sports          0.118446\n",
      "cute            0.741213\n",
      "sex             0.142406\n",
      "sexy            0.094746\n",
      "hot             0.241635\n",
      "kissed          0.071102\n",
      "dance           0.812517\n",
      "band            0.230881\n",
      "marching        0.017488\n",
      "music           0.858772\n",
      "rock            0.220353\n",
      "god             0.404753\n",
      "church          0.256939\n",
      "jesus           0.066069\n",
      "bible           0.004487\n",
      "hair            0.620815\n",
      "dress           0.400691\n",
      "blonde          0.114677\n",
      "mall            0.775669\n",
      "shopping        1.438070\n",
      "clothes         0.408343\n",
      "hollister       0.069347\n",
      "abercrombie     0.035091\n",
      "die             0.166422\n",
      "death           0.123910\n",
      "drunk           0.082089\n",
      "drugs           0.032963\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 6\n",
      "Element count 493\n",
      "basketball      0.816874\n",
      "football        0.939873\n",
      "soccer          0.367084\n",
      "softball        0.220208\n",
      "volleyball      0.123641\n",
      "swimming        0.179279\n",
      "cheerleading    0.208909\n",
      "baseball        3.201824\n",
      "tennis          0.094649\n",
      "sports          0.437002\n",
      "cute            0.399913\n",
      "sex             0.280323\n",
      "sexy            0.146664\n",
      "hot             0.163078\n",
      "kissed          0.095076\n",
      "dance           0.412705\n",
      "band            0.209978\n",
      "marching        0.015988\n",
      "music           0.811979\n",
      "rock            0.593684\n",
      "god             0.445534\n",
      "church          0.364739\n",
      "jesus           0.091878\n",
      "bible           0.015349\n",
      "hair            0.510764\n",
      "dress           0.043062\n",
      "blonde          0.330203\n",
      "mall            0.289916\n",
      "shopping        0.201452\n",
      "clothes         0.129397\n",
      "hollister       0.029206\n",
      "abercrombie     0.022597\n",
      "die             0.146878\n",
      "death           0.106374\n",
      "drunk           0.054573\n",
      "drugs           0.053294\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 7\n",
      "Element count 889\n",
      "basketball      0.500166\n",
      "football        0.539076\n",
      "soccer          0.323992\n",
      "softball        0.238229\n",
      "volleyball      0.195915\n",
      "swimming        0.278387\n",
      "cheerleading    0.250026\n",
      "baseball        0.188201\n",
      "tennis          0.156777\n",
      "sports          0.493811\n",
      "cute            0.713890\n",
      "sex             2.600641\n",
      "sexy            0.324898\n",
      "hot             0.270673\n",
      "kissed          1.790670\n",
      "dance           0.930905\n",
      "band            0.757565\n",
      "marching        0.056721\n",
      "music           2.041725\n",
      "rock            0.929087\n",
      "god             0.924099\n",
      "church          0.364037\n",
      "jesus           0.143959\n",
      "bible           0.034373\n",
      "hair            3.283219\n",
      "dress           0.373904\n",
      "blonde          0.888019\n",
      "mall            0.701184\n",
      "shopping        0.544523\n",
      "clothes         0.737937\n",
      "hollister       0.133181\n",
      "abercrombie     0.135563\n",
      "die             1.305597\n",
      "death           0.534536\n",
      "drunk           0.937365\n",
      "drugs           1.094935\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 8\n",
      "Element count 859\n",
      "basketball      0.385742\n",
      "football        0.407691\n",
      "soccer          0.322434\n",
      "softball        0.164279\n",
      "volleyball      0.256583\n",
      "swimming        0.244684\n",
      "cheerleading    0.313190\n",
      "baseball        0.103397\n",
      "tennis          0.150415\n",
      "sports          0.179412\n",
      "cute            0.620374\n",
      "sex             0.240411\n",
      "sexy            0.192351\n",
      "hot             0.334331\n",
      "kissed          0.144292\n",
      "dance           0.687958\n",
      "band            0.205755\n",
      "marching        0.014903\n",
      "music           0.905612\n",
      "rock            0.296210\n",
      "god             0.467999\n",
      "church          0.231632\n",
      "jesus           0.113332\n",
      "bible           0.015827\n",
      "hair            0.939917\n",
      "dress           0.173636\n",
      "blonde          0.214416\n",
      "mall            0.657456\n",
      "shopping        0.904566\n",
      "clothes         0.404917\n",
      "hollister       1.493391\n",
      "abercrombie     1.158369\n",
      "die             0.219731\n",
      "death           0.158618\n",
      "drunk           0.097158\n",
      "drugs           0.078789\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 9\n",
      "Element count 1664\n",
      "basketball      0.268228\n",
      "football        0.277146\n",
      "soccer          0.267411\n",
      "softball        0.116750\n",
      "volleyball      0.097533\n",
      "swimming        0.147836\n",
      "cheerleading    0.102242\n",
      "baseball        0.087170\n",
      "tennis          0.086479\n",
      "sports          0.138919\n",
      "cute            0.856048\n",
      "sex             0.469882\n",
      "sexy            1.135448\n",
      "hot             0.453489\n",
      "kissed          0.131005\n",
      "dance           1.185631\n",
      "band            1.650232\n",
      "marching        0.371658\n",
      "music           1.934477\n",
      "rock            0.996094\n",
      "god             0.544745\n",
      "church          0.221692\n",
      "jesus           0.087296\n",
      "bible           0.011556\n",
      "hair            0.766433\n",
      "dress           0.138604\n",
      "blonde          0.207685\n",
      "mall            0.293411\n",
      "shopping        0.294167\n",
      "clothes         0.235758\n",
      "hollister       0.031402\n",
      "abercrombie     0.017334\n",
      "die             0.288575\n",
      "death           0.195251\n",
      "drunk           0.122966\n",
      "drugs           0.057590\n",
      "dtype: float64\n",
      "==========================\n"
     ]
    }
   ],
   "source": [
    "for i in xrange(9):\n",
    "    print \"Centroid \" + str(i+1)\n",
    "    print \"Element count \" + str(sum([1 for  j in km.labels if j == i]))\n",
    "    print km.centroids[i] * df_sns_words.values.std(axis=0) + df_sns_words.mean(axis=0)\n",
    "    print \"==========================\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "## KMeans с MiniBatchKMeans\n",
    "km = MiniBatchKMeans(9, max_iter=100, batch_size=df_sns_words1.shape[0]/10)\n",
    "km.fit(df_sns_words1.values)\n",
    "km.labels = km.predict(df_sns_words1.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Centroid 1\n",
      "Element count 1418\n",
      "basketball      1.880255\n",
      "football        0.649745\n",
      "soccer          0.471896\n",
      "softball        1.762056\n",
      "volleyball      1.248601\n",
      "swimming        0.156724\n",
      "cheerleading    0.140568\n",
      "baseball        0.171222\n",
      "tennis          0.132974\n",
      "sports          0.698969\n",
      "cute            0.308755\n",
      "sex             0.151340\n",
      "sexy            0.097142\n",
      "hot             0.150165\n",
      "kissed          0.077741\n",
      "dance           0.486881\n",
      "band            0.241164\n",
      "marching        0.026857\n",
      "music           0.761184\n",
      "rock            0.247376\n",
      "god             0.445872\n",
      "church          0.288663\n",
      "jesus           0.094864\n",
      "bible           0.011047\n",
      "hair            0.472313\n",
      "dress           0.110812\n",
      "blonde          0.122894\n",
      "mall            0.275476\n",
      "shopping        0.427643\n",
      "clothes         0.192902\n",
      "hollister       0.039078\n",
      "abercrombie     0.028446\n",
      "die             0.143262\n",
      "death           0.121099\n",
      "drunk           0.063312\n",
      "drugs           0.036592\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 2\n",
      "Element count 2389\n",
      "basketball      0.228717\n",
      "football        0.878007\n",
      "soccer          0.276386\n",
      "softball        0.088004\n",
      "volleyball      0.103497\n",
      "swimming        0.133049\n",
      "cheerleading    0.101898\n",
      "baseball        0.076116\n",
      "tennis          0.233715\n",
      "sports          0.145878\n",
      "cute            0.285036\n",
      "sex             0.132313\n",
      "sexy            0.057345\n",
      "hot             0.083822\n",
      "kissed          0.055008\n",
      "dance           0.468868\n",
      "band            0.469276\n",
      "marching        0.068410\n",
      "music           0.887689\n",
      "rock            0.259335\n",
      "god             1.656536\n",
      "church          1.027122\n",
      "jesus           0.605271\n",
      "bible           0.133006\n",
      "hair            0.383367\n",
      "dress           0.104029\n",
      "blonde          0.084109\n",
      "mall            0.205969\n",
      "shopping        0.235031\n",
      "clothes         0.113252\n",
      "hollister       0.023282\n",
      "abercrombie     0.012584\n",
      "die             0.227363\n",
      "death           0.176742\n",
      "drunk           0.061074\n",
      "drugs           0.033488\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 3\n",
      "Element count 18385\n",
      "basketball      0.120851\n",
      "football        0.098591\n",
      "soccer          0.136345\n",
      "softball        0.056023\n",
      "volleyball      0.058009\n",
      "swimming        0.066911\n",
      "cheerleading    0.046480\n",
      "baseball        0.030571\n",
      "tennis          0.052416\n",
      "sports          0.072352\n",
      "cute            0.165495\n",
      "sex             0.094995\n",
      "sexy            0.064903\n",
      "hot             0.060549\n",
      "kissed          0.034026\n",
      "dance           0.239719\n",
      "band            0.163396\n",
      "marching        0.015689\n",
      "music           0.513844\n",
      "rock            0.135602\n",
      "god             0.232045\n",
      "church          0.070858\n",
      "jesus           0.033036\n",
      "bible           0.000883\n",
      "hair            0.189344\n",
      "dress           0.040383\n",
      "blonde          0.040490\n",
      "mall            0.117595\n",
      "shopping        0.126143\n",
      "clothes         0.053284\n",
      "hollister       0.014199\n",
      "abercrombie     0.009252\n",
      "die             0.119606\n",
      "death           0.073466\n",
      "drunk           0.053676\n",
      "drugs           0.021281\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 4\n",
      "Element count 311\n",
      "basketball      0.392043\n",
      "football        0.313717\n",
      "soccer          0.239448\n",
      "softball        0.244723\n",
      "volleyball      0.133117\n",
      "swimming        0.183441\n",
      "cheerleading    0.237417\n",
      "baseball        0.112419\n",
      "tennis          0.124594\n",
      "sports          0.189123\n",
      "cute            0.440745\n",
      "sex             0.185065\n",
      "sexy            0.134335\n",
      "hot             0.189934\n",
      "kissed          0.127840\n",
      "dance           0.538147\n",
      "band            0.472400\n",
      "marching        0.045049\n",
      "music           1.058030\n",
      "rock            0.323456\n",
      "god             4.803093\n",
      "church          4.973136\n",
      "jesus           1.900538\n",
      "bible           0.732943\n",
      "hair            0.538147\n",
      "dress           0.195615\n",
      "blonde          0.108766\n",
      "mall            0.424104\n",
      "shopping        0.665985\n",
      "clothes         0.327107\n",
      "hollister       0.063718\n",
      "abercrombie     0.019481\n",
      "die             0.387578\n",
      "death           0.245939\n",
      "drunk           0.109983\n",
      "drugs           0.076298\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 5\n",
      "Element count 3592\n",
      "basketball      0.236058\n",
      "football        0.255760\n",
      "soccer          0.441478\n",
      "softball        0.099291\n",
      "volleyball      0.141887\n",
      "swimming        0.413922\n",
      "cheerleading    0.318229\n",
      "baseball        0.061525\n",
      "tennis          0.120142\n",
      "sports          0.118446\n",
      "cute            0.741213\n",
      "sex             0.142406\n",
      "sexy            0.094746\n",
      "hot             0.241635\n",
      "kissed          0.071102\n",
      "dance           0.812517\n",
      "band            0.230881\n",
      "marching        0.017488\n",
      "music           0.858772\n",
      "rock            0.220353\n",
      "god             0.404753\n",
      "church          0.256939\n",
      "jesus           0.066069\n",
      "bible           0.004487\n",
      "hair            0.620815\n",
      "dress           0.400691\n",
      "blonde          0.114677\n",
      "mall            0.775669\n",
      "shopping        1.438070\n",
      "clothes         0.408343\n",
      "hollister       0.069347\n",
      "abercrombie     0.035091\n",
      "die             0.166422\n",
      "death           0.123910\n",
      "drunk           0.082089\n",
      "drugs           0.032963\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 6\n",
      "Element count 493\n",
      "basketball      0.816874\n",
      "football        0.939873\n",
      "soccer          0.367084\n",
      "softball        0.220208\n",
      "volleyball      0.123641\n",
      "swimming        0.179279\n",
      "cheerleading    0.208909\n",
      "baseball        3.201824\n",
      "tennis          0.094649\n",
      "sports          0.437002\n",
      "cute            0.399913\n",
      "sex             0.280323\n",
      "sexy            0.146664\n",
      "hot             0.163078\n",
      "kissed          0.095076\n",
      "dance           0.412705\n",
      "band            0.209978\n",
      "marching        0.015988\n",
      "music           0.811979\n",
      "rock            0.593684\n",
      "god             0.445534\n",
      "church          0.364739\n",
      "jesus           0.091878\n",
      "bible           0.015349\n",
      "hair            0.510764\n",
      "dress           0.043062\n",
      "blonde          0.330203\n",
      "mall            0.289916\n",
      "shopping        0.201452\n",
      "clothes         0.129397\n",
      "hollister       0.029206\n",
      "abercrombie     0.022597\n",
      "die             0.146878\n",
      "death           0.106374\n",
      "drunk           0.054573\n",
      "drugs           0.053294\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 7\n",
      "Element count 889\n",
      "basketball      0.500166\n",
      "football        0.539076\n",
      "soccer          0.323992\n",
      "softball        0.238229\n",
      "volleyball      0.195915\n",
      "swimming        0.278387\n",
      "cheerleading    0.250026\n",
      "baseball        0.188201\n",
      "tennis          0.156777\n",
      "sports          0.493811\n",
      "cute            0.713890\n",
      "sex             2.600641\n",
      "sexy            0.324898\n",
      "hot             0.270673\n",
      "kissed          1.790670\n",
      "dance           0.930905\n",
      "band            0.757565\n",
      "marching        0.056721\n",
      "music           2.041725\n",
      "rock            0.929087\n",
      "god             0.924099\n",
      "church          0.364037\n",
      "jesus           0.143959\n",
      "bible           0.034373\n",
      "hair            3.283219\n",
      "dress           0.373904\n",
      "blonde          0.888019\n",
      "mall            0.701184\n",
      "shopping        0.544523\n",
      "clothes         0.737937\n",
      "hollister       0.133181\n",
      "abercrombie     0.135563\n",
      "die             1.305597\n",
      "death           0.534536\n",
      "drunk           0.937365\n",
      "drugs           1.094935\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 8\n",
      "Element count 859\n",
      "basketball      0.385742\n",
      "football        0.407691\n",
      "soccer          0.322434\n",
      "softball        0.164279\n",
      "volleyball      0.256583\n",
      "swimming        0.244684\n",
      "cheerleading    0.313190\n",
      "baseball        0.103397\n",
      "tennis          0.150415\n",
      "sports          0.179412\n",
      "cute            0.620374\n",
      "sex             0.240411\n",
      "sexy            0.192351\n",
      "hot             0.334331\n",
      "kissed          0.144292\n",
      "dance           0.687958\n",
      "band            0.205755\n",
      "marching        0.014903\n",
      "music           0.905612\n",
      "rock            0.296210\n",
      "god             0.467999\n",
      "church          0.231632\n",
      "jesus           0.113332\n",
      "bible           0.015827\n",
      "hair            0.939917\n",
      "dress           0.173636\n",
      "blonde          0.214416\n",
      "mall            0.657456\n",
      "shopping        0.904566\n",
      "clothes         0.404917\n",
      "hollister       1.493391\n",
      "abercrombie     1.158369\n",
      "die             0.219731\n",
      "death           0.158618\n",
      "drunk           0.097158\n",
      "drugs           0.078789\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 9\n",
      "Element count 1664\n",
      "basketball      0.268228\n",
      "football        0.277146\n",
      "soccer          0.267411\n",
      "softball        0.116750\n",
      "volleyball      0.097533\n",
      "swimming        0.147836\n",
      "cheerleading    0.102242\n",
      "baseball        0.087170\n",
      "tennis          0.086479\n",
      "sports          0.138919\n",
      "cute            0.856048\n",
      "sex             0.469882\n",
      "sexy            1.135448\n",
      "hot             0.453489\n",
      "kissed          0.131005\n",
      "dance           1.185631\n",
      "band            1.650232\n",
      "marching        0.371658\n",
      "music           1.934477\n",
      "rock            0.996094\n",
      "god             0.544745\n",
      "church          0.221692\n",
      "jesus           0.087296\n",
      "bible           0.011556\n",
      "hair            0.766433\n",
      "dress           0.138604\n",
      "blonde          0.207685\n",
      "mall            0.293411\n",
      "shopping        0.294167\n",
      "clothes         0.235758\n",
      "hollister       0.031402\n",
      "abercrombie     0.017334\n",
      "die             0.288575\n",
      "death           0.195251\n",
      "drunk           0.122966\n",
      "drugs           0.057590\n",
      "dtype: float64\n",
      "==========================\n"
     ]
    }
   ],
   "source": [
    "for i in xrange(9):\n",
    "    print \"Centroid \" + str(i+1)\n",
    "    print \"Element count \" + str(sum([1 for  j in km.labels if j == i]))\n",
    "    print km.centroids[i] * df_sns_words.values.std(axis=0) + df_sns_words.mean(axis=0)\n",
    "    print \"==========================\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "## KMeans++ с MiniBatchKMeans\n",
    "km = MiniBatchKMeans(9, init=\"k-means\", max_iter=100, batch_size=df_sns_words1.shape[0]/10)\n",
    "km.fit(df_sns_words1.values)\n",
    "km.labels = km.predict(df_sns_words1.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Centroid 1\n",
      "Element count 1418\n",
      "basketball      1.880255\n",
      "football        0.649745\n",
      "soccer          0.471896\n",
      "softball        1.762056\n",
      "volleyball      1.248601\n",
      "swimming        0.156724\n",
      "cheerleading    0.140568\n",
      "baseball        0.171222\n",
      "tennis          0.132974\n",
      "sports          0.698969\n",
      "cute            0.308755\n",
      "sex             0.151340\n",
      "sexy            0.097142\n",
      "hot             0.150165\n",
      "kissed          0.077741\n",
      "dance           0.486881\n",
      "band            0.241164\n",
      "marching        0.026857\n",
      "music           0.761184\n",
      "rock            0.247376\n",
      "god             0.445872\n",
      "church          0.288663\n",
      "jesus           0.094864\n",
      "bible           0.011047\n",
      "hair            0.472313\n",
      "dress           0.110812\n",
      "blonde          0.122894\n",
      "mall            0.275476\n",
      "shopping        0.427643\n",
      "clothes         0.192902\n",
      "hollister       0.039078\n",
      "abercrombie     0.028446\n",
      "die             0.143262\n",
      "death           0.121099\n",
      "drunk           0.063312\n",
      "drugs           0.036592\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 2\n",
      "Element count 2389\n",
      "basketball      0.228717\n",
      "football        0.878007\n",
      "soccer          0.276386\n",
      "softball        0.088004\n",
      "volleyball      0.103497\n",
      "swimming        0.133049\n",
      "cheerleading    0.101898\n",
      "baseball        0.076116\n",
      "tennis          0.233715\n",
      "sports          0.145878\n",
      "cute            0.285036\n",
      "sex             0.132313\n",
      "sexy            0.057345\n",
      "hot             0.083822\n",
      "kissed          0.055008\n",
      "dance           0.468868\n",
      "band            0.469276\n",
      "marching        0.068410\n",
      "music           0.887689\n",
      "rock            0.259335\n",
      "god             1.656536\n",
      "church          1.027122\n",
      "jesus           0.605271\n",
      "bible           0.133006\n",
      "hair            0.383367\n",
      "dress           0.104029\n",
      "blonde          0.084109\n",
      "mall            0.205969\n",
      "shopping        0.235031\n",
      "clothes         0.113252\n",
      "hollister       0.023282\n",
      "abercrombie     0.012584\n",
      "die             0.227363\n",
      "death           0.176742\n",
      "drunk           0.061074\n",
      "drugs           0.033488\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 3\n",
      "Element count 18385\n",
      "basketball      0.120851\n",
      "football        0.098591\n",
      "soccer          0.136345\n",
      "softball        0.056023\n",
      "volleyball      0.058009\n",
      "swimming        0.066911\n",
      "cheerleading    0.046480\n",
      "baseball        0.030571\n",
      "tennis          0.052416\n",
      "sports          0.072352\n",
      "cute            0.165495\n",
      "sex             0.094995\n",
      "sexy            0.064903\n",
      "hot             0.060549\n",
      "kissed          0.034026\n",
      "dance           0.239719\n",
      "band            0.163396\n",
      "marching        0.015689\n",
      "music           0.513844\n",
      "rock            0.135602\n",
      "god             0.232045\n",
      "church          0.070858\n",
      "jesus           0.033036\n",
      "bible           0.000883\n",
      "hair            0.189344\n",
      "dress           0.040383\n",
      "blonde          0.040490\n",
      "mall            0.117595\n",
      "shopping        0.126143\n",
      "clothes         0.053284\n",
      "hollister       0.014199\n",
      "abercrombie     0.009252\n",
      "die             0.119606\n",
      "death           0.073466\n",
      "drunk           0.053676\n",
      "drugs           0.021281\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 4\n",
      "Element count 311\n",
      "basketball      0.392043\n",
      "football        0.313717\n",
      "soccer          0.239448\n",
      "softball        0.244723\n",
      "volleyball      0.133117\n",
      "swimming        0.183441\n",
      "cheerleading    0.237417\n",
      "baseball        0.112419\n",
      "tennis          0.124594\n",
      "sports          0.189123\n",
      "cute            0.440745\n",
      "sex             0.185065\n",
      "sexy            0.134335\n",
      "hot             0.189934\n",
      "kissed          0.127840\n",
      "dance           0.538147\n",
      "band            0.472400\n",
      "marching        0.045049\n",
      "music           1.058030\n",
      "rock            0.323456\n",
      "god             4.803093\n",
      "church          4.973136\n",
      "jesus           1.900538\n",
      "bible           0.732943\n",
      "hair            0.538147\n",
      "dress           0.195615\n",
      "blonde          0.108766\n",
      "mall            0.424104\n",
      "shopping        0.665985\n",
      "clothes         0.327107\n",
      "hollister       0.063718\n",
      "abercrombie     0.019481\n",
      "die             0.387578\n",
      "death           0.245939\n",
      "drunk           0.109983\n",
      "drugs           0.076298\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 5\n",
      "Element count 3592\n",
      "basketball      0.236058\n",
      "football        0.255760\n",
      "soccer          0.441478\n",
      "softball        0.099291\n",
      "volleyball      0.141887\n",
      "swimming        0.413922\n",
      "cheerleading    0.318229\n",
      "baseball        0.061525\n",
      "tennis          0.120142\n",
      "sports          0.118446\n",
      "cute            0.741213\n",
      "sex             0.142406\n",
      "sexy            0.094746\n",
      "hot             0.241635\n",
      "kissed          0.071102\n",
      "dance           0.812517\n",
      "band            0.230881\n",
      "marching        0.017488\n",
      "music           0.858772\n",
      "rock            0.220353\n",
      "god             0.404753\n",
      "church          0.256939\n",
      "jesus           0.066069\n",
      "bible           0.004487\n",
      "hair            0.620815\n",
      "dress           0.400691\n",
      "blonde          0.114677\n",
      "mall            0.775669\n",
      "shopping        1.438070\n",
      "clothes         0.408343\n",
      "hollister       0.069347\n",
      "abercrombie     0.035091\n",
      "die             0.166422\n",
      "death           0.123910\n",
      "drunk           0.082089\n",
      "drugs           0.032963\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 6\n",
      "Element count 493\n",
      "basketball      0.816874\n",
      "football        0.939873\n",
      "soccer          0.367084\n",
      "softball        0.220208\n",
      "volleyball      0.123641\n",
      "swimming        0.179279\n",
      "cheerleading    0.208909\n",
      "baseball        3.201824\n",
      "tennis          0.094649\n",
      "sports          0.437002\n",
      "cute            0.399913\n",
      "sex             0.280323\n",
      "sexy            0.146664\n",
      "hot             0.163078\n",
      "kissed          0.095076\n",
      "dance           0.412705\n",
      "band            0.209978\n",
      "marching        0.015988\n",
      "music           0.811979\n",
      "rock            0.593684\n",
      "god             0.445534\n",
      "church          0.364739\n",
      "jesus           0.091878\n",
      "bible           0.015349\n",
      "hair            0.510764\n",
      "dress           0.043062\n",
      "blonde          0.330203\n",
      "mall            0.289916\n",
      "shopping        0.201452\n",
      "clothes         0.129397\n",
      "hollister       0.029206\n",
      "abercrombie     0.022597\n",
      "die             0.146878\n",
      "death           0.106374\n",
      "drunk           0.054573\n",
      "drugs           0.053294\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 7\n",
      "Element count 889\n",
      "basketball      0.500166\n",
      "football        0.539076\n",
      "soccer          0.323992\n",
      "softball        0.238229\n",
      "volleyball      0.195915\n",
      "swimming        0.278387\n",
      "cheerleading    0.250026\n",
      "baseball        0.188201\n",
      "tennis          0.156777\n",
      "sports          0.493811\n",
      "cute            0.713890\n",
      "sex             2.600641\n",
      "sexy            0.324898\n",
      "hot             0.270673\n",
      "kissed          1.790670\n",
      "dance           0.930905\n",
      "band            0.757565\n",
      "marching        0.056721\n",
      "music           2.041725\n",
      "rock            0.929087\n",
      "god             0.924099\n",
      "church          0.364037\n",
      "jesus           0.143959\n",
      "bible           0.034373\n",
      "hair            3.283219\n",
      "dress           0.373904\n",
      "blonde          0.888019\n",
      "mall            0.701184\n",
      "shopping        0.544523\n",
      "clothes         0.737937\n",
      "hollister       0.133181\n",
      "abercrombie     0.135563\n",
      "die             1.305597\n",
      "death           0.534536\n",
      "drunk           0.937365\n",
      "drugs           1.094935\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 8\n",
      "Element count 859\n",
      "basketball      0.385742\n",
      "football        0.407691\n",
      "soccer          0.322434\n",
      "softball        0.164279\n",
      "volleyball      0.256583\n",
      "swimming        0.244684\n",
      "cheerleading    0.313190\n",
      "baseball        0.103397\n",
      "tennis          0.150415\n",
      "sports          0.179412\n",
      "cute            0.620374\n",
      "sex             0.240411\n",
      "sexy            0.192351\n",
      "hot             0.334331\n",
      "kissed          0.144292\n",
      "dance           0.687958\n",
      "band            0.205755\n",
      "marching        0.014903\n",
      "music           0.905612\n",
      "rock            0.296210\n",
      "god             0.467999\n",
      "church          0.231632\n",
      "jesus           0.113332\n",
      "bible           0.015827\n",
      "hair            0.939917\n",
      "dress           0.173636\n",
      "blonde          0.214416\n",
      "mall            0.657456\n",
      "shopping        0.904566\n",
      "clothes         0.404917\n",
      "hollister       1.493391\n",
      "abercrombie     1.158369\n",
      "die             0.219731\n",
      "death           0.158618\n",
      "drunk           0.097158\n",
      "drugs           0.078789\n",
      "dtype: float64\n",
      "==========================\n",
      "Centroid 9\n",
      "Element count 1664\n",
      "basketball      0.268228\n",
      "football        0.277146\n",
      "soccer          0.267411\n",
      "softball        0.116750\n",
      "volleyball      0.097533\n",
      "swimming        0.147836\n",
      "cheerleading    0.102242\n",
      "baseball        0.087170\n",
      "tennis          0.086479\n",
      "sports          0.138919\n",
      "cute            0.856048\n",
      "sex             0.469882\n",
      "sexy            1.135448\n",
      "hot             0.453489\n",
      "kissed          0.131005\n",
      "dance           1.185631\n",
      "band            1.650232\n",
      "marching        0.371658\n",
      "music           1.934477\n",
      "rock            0.996094\n",
      "god             0.544745\n",
      "church          0.221692\n",
      "jesus           0.087296\n",
      "bible           0.011556\n",
      "hair            0.766433\n",
      "dress           0.138604\n",
      "blonde          0.207685\n",
      "mall            0.293411\n",
      "shopping        0.294167\n",
      "clothes         0.235758\n",
      "hollister       0.031402\n",
      "abercrombie     0.017334\n",
      "die             0.288575\n",
      "death           0.195251\n",
      "drunk           0.122966\n",
      "drugs           0.057590\n",
      "dtype: float64\n",
      "==========================\n"
     ]
    }
   ],
   "source": [
    "for i in xrange(9):\n",
    "    print \"Centroid \" + str(i+1)\n",
    "    print \"Element count \" + str(sum([1 for  j in km.labels if j == i]))\n",
    "    print km.centroids[i] * df_sns_words.values.std(axis=0) + df_sns_words.mean(axis=0)\n",
    "    print \"==========================\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Вывод\n",
    "### Kmeans\n",
    "1. преобладают слова волейболисты, женский бейсбол, баскетбол. Скорее всего это спортсмены\n",
    "2. преобладают слова Бог и церковь, скорее всего религиозные люди\n",
    "3. серая масса, неинтерпретируемый кластер\n",
    "4. очень сильно преобладают религиозные слова, скорее всего сильно религиозные люди\n",
    "5. преобладает слово шоппинг, Малл, милый, скорее всего модницы\n",
    "6. бейсболисты\n",
    "7. секс, наркотики, рок-н-ролл\n",
    "8. abercrombie и hollister - похоже на названия магазинов одежды\n",
    "9. танцы, тусовки, музыка\n",
    "\n",
    "### Остальные алгоритмы\n",
    "Остальные алгоритмы дают примерно такой же результат. Т.к. данных весьма много и кластеры чётко выделяются."
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "2.7.13"
  },
  "nav_menu": {},
  "toc": {
   "navigate_menu": true,
   "number_sections": false,
   "sideBar": true,
   "threshold": 6,
   "toc_cell": false,
   "toc_section_display": "block",
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}