Временные_ряды_4.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Домашнее задание\n",
    "\n",
    "- Скачайте текст войны и мир (ссылка на [github](https://github.com/netology-code/ds3-spring-2018/tree/master/7.%20Time_series/7.4/war_and_peace) и [гугл диск](https://drive.google.com/drive/u/2/folders/1aM7cNI6Qs7AyeYpQ8lNg7vLsHzIl0JbU)). Очистите его, оставив только русские буквы и символы пробела. Постройте матрицу переходных вероятностей. Посчитайте вероятность встретить строку из 4-х символов, которая начинается п, а кончается на р; Используя матрицу переходных вероятностей сгенерируйте слово, начинающееся с буквы \"Б\" и имеющее длину 8-мь символов\n",
    "- [Выберите ряд](https://datamarket.com/data/list/?q=cat:fpg%20provider:tsdl) и оцените его через HMM. Попробуйте \"построить\" (через среднее каждой компоненты) наивный прогноз ряда."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import re\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "%matplotlib inline\n",
    "sns.set()\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Часть 1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Debugging support not compiled in. Ignoring --debug flag.\n",
      "--2020-05-11 13:49:32--  https://raw.githubusercontent.com/netology-code/ds3-spring-2018/master/7.%20Time_series/7.4/war_and_peace/tom_1.txt\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.244.133\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.244.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1273582 (1,2M) [text/plain]\n",
      "Saving to: ‘tom_1.txt’\n",
      "\n",
      "tom_1.txt           100%[===================>]   1,21M  2,30MB/s    in 0,5s    \n",
      "\n",
      "2020-05-11 13:49:32 (2,30 MB/s) - ‘tom_1.txt’ saved [1273582/1273582]\n",
      "\n",
      "Debugging support not compiled in. Ignoring --debug flag.\n",
      "--2020-05-11 13:49:32--  https://raw.githubusercontent.com/netology-code/ds3-spring-2018/master/7.%20Time_series/7.4/war_and_peace/tom_2.txt\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.244.133\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.244.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1356802 (1,3M) [text/plain]\n",
      "Saving to: ‘tom_2.txt’\n",
      "\n",
      "tom_2.txt           100%[===================>]   1,29M  2,28MB/s    in 0,6s    \n",
      "\n",
      "2020-05-11 13:49:33 (2,28 MB/s) - ‘tom_2.txt’ saved [1356802/1356802]\n",
      "\n",
      "Debugging support not compiled in. Ignoring --debug flag.\n",
      "--2020-05-11 13:49:33--  https://raw.githubusercontent.com/netology-code/ds3-spring-2018/master/7.%20Time_series/7.4/war_and_peace/tom_3.txt\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.244.133\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.244.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1467361 (1,4M) [text/plain]\n",
      "Saving to: ‘tom_3.txt’\n",
      "\n",
      "tom_3.txt           100%[===================>]   1,40M  2,08MB/s    in 0,7s    \n",
      "\n",
      "2020-05-11 13:49:34 (2,08 MB/s) - ‘tom_3.txt’ saved [1467361/1467361]\n",
      "\n",
      "Debugging support not compiled in. Ignoring --debug flag.\n",
      "--2020-05-11 13:49:34--  https://raw.githubusercontent.com/netology-code/ds3-spring-2018/master/7.%20Time_series/7.4/war_and_peace/tom_4.txt\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.244.133\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.244.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1257579 (1,2M) [text/plain]\n",
      "Saving to: ‘tom_4.txt’\n",
      "\n",
      "tom_4.txt           100%[===================>]   1,20M  2,26MB/s    in 0,5s    \n",
      "\n",
      "2020-05-11 13:49:35 (2,26 MB/s) - ‘tom_4.txt’ saved [1257579/1257579]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "!wget -d 'https://raw.githubusercontent.com/netology-code/ds3-spring-2018/master/7.%20Time_series/7.4/war_and_peace/tom_1.txt'\n",
    "!wget -d 'https://raw.githubusercontent.com/netology-code/ds3-spring-2018/master/7.%20Time_series/7.4/war_and_peace/tom_2.txt'\n",
    "!wget -d 'https://raw.githubusercontent.com/netology-code/ds3-spring-2018/master/7.%20Time_series/7.4/war_and_peace/tom_3.txt'\n",
    "!wget -d 'https://raw.githubusercontent.com/netology-code/ds3-spring-2018/master/7.%20Time_series/7.4/war_and_peace/tom_4.txt'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "text = ''\n",
    "for i in range(4):\n",
    "    with open('tom_{}.txt'.format(i+1), 'r') as f:\n",
    "        text += f.read().lower()\n",
    "text = re.sub('[^а-я\\s]','', text)\n",
    "text = re.sub('\\s',' ', text)\n",
    "text = re.sub(' +',' ', text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "cyr = 'абвгдежзийклмнопрстуфхцчшщъыьэюя '\n",
    "pos = {j:i for i,j in enumerate(cyr)}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "probs = np.zeros((len(cyr), len(cyr)))\n",
    "for cp, cn in zip(text[:-1], text[1:]):\n",
    "    probs[pos[cp], pos[cn]] += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "probs = probs / probs.sum(axis=1)[:, None]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(probs, index = (i for i in cyr), columns = (i for i in cyr))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x720 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(16, 10))\n",
    "sns.heatmap(df)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "start = pos['п']\n",
    "end = pos['р']\n",
    "length = 4\n",
    "prior = 1. / len(cyr)\n",
    "for i in range(length):\n",
    "    index = pos[s[i]]\n",
    "    next_char = np.random.choice(list(cyr), p=probs[index])\n",
    "    s.append(next_char)\n",
    "\n",
    "print (''.join(s))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Вероятность встретить слово \"п**р\" равняется 0.03128474210545203.\n"
     ]
    }
   ],
   "source": [
    "start = pos['п']\n",
    "end = pos['р']\n",
    "prior = 1. / len(cyr)\n",
    "proba_fin = 0\n",
    "proba_fin += prior\n",
    "for c1 in cyr:\n",
    "    for c2 in cyr:\n",
    "        proba = prior * probs[start, pos[c1]] * probs[pos[c1], pos[c2]] * probs[pos[c2], end]\n",
    "        if proba > 0:\n",
    "            proba_fin += proba\n",
    "print('Вероятность встретить слово \"п**р\" равняется {}.'.format(proba_fin))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "бытеноня \n"
     ]
    }
   ],
   "source": [
    "s = []\n",
    "start = 'б'\n",
    "s.append(start)\n",
    "length = 8\n",
    "for i in range(length):\n",
    "    index = pos[s[i]]\n",
    "    next_char = np.random.choice(list(cyr), p=probs[index])\n",
    "    s.append(next_char)\n",
    "\n",
    "print (''.join(s))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Часть 2."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "С ссылкой проблемы, поэтому выберу ряд из прошлых лекций."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "from hmmlearn.hmm import GaussianHMM\n",
    "series = pd.read_csv('../Архив/Series/international-airline-passengers.csv')[\"Count\"]\n",
    "values = series.values.reshape(-1,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 -893.1837902090575\n",
      "2 -803.8479638762091\n",
      "3 -767.0819417008181\n",
      "4 -743.9977519551946\n",
      "5 -731.3629628719801\n",
      "6 -706.6057596024548\n",
      "7 -694.9945555321286\n",
      "8 -673.9336800691148\n",
      "9 -664.2465557766129\n",
      "10 -656.1684316480492\n",
      "11 -650.2620800745818\n"
     ]
    }
   ],
   "source": [
    "for n in range(1, 12):\n",
    "    model = GaussianHMM(n_components=n,\n",
    "                        covariance_type=\"diag\",\n",
    "                        n_iter=1000)\n",
    "    model.fit(values)\n",
    "    print (n, model.score(values))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1278a5d68>]"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = GaussianHMM(n_components=11,\n",
    "                        covariance_type=\"diag\",\n",
    "                        n_iter=1000)\n",
    "model.fit(values)\n",
    "labels = model.predict(values)\n",
    "means = np.zeros_like(values)\n",
    "for i in range(model.n_components):\n",
    "    means[labels == i] = model.means_[i]\n",
    "    \n",
    "plt.figure(figsize=(14, 5))\n",
    "plt.plot(values, linewidth=2)\n",
    "plt.plot(means, linewidth=3, alpha=.4, color='red')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}