NLP2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Cross-Entropie a modelování jazyka"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Rozdělení na datasety:\n",
    "* posledních 20 000 testovací data\n",
    "* posledních 40 000 ze zbytku heldout data\n",
    "* zbytek trénovací data\n",
    "\n",
    "### Spočítat pravděpodobnosti\n",
    "$T$ velikost textu, $V$ velikost slovníku  \n",
    "$\n",
    "p_0(w_i) = \\dfrac{1}{V}\\\\\n",
    "p_1(w_i) = \\dfrac{c_1(w_i)}{T}\\\\\n",
    "p_2(w_i|w_{i-1}) = \\dfrac{c_2(w_{i-1},w_i)}{c_1({w_{i-1})}}\\\\\n",
    "p_3(w_i|w_{i-2},w_{i-1}) = \\dfrac{c_3(w_{i-2},w_{i-1},w_i)}{c_2(w_{i-1},w_i)}\n",
    "$\n",
    "\n",
    "### EM algoritmus\n",
    "Vypočítat parametry z heldout dat (také z trénovacích a poté zahodit)\n",
    "$$p_s(w_i|w_{i-2},w_{i-1}) = l_0p_0(w_i) + l_1p_1(w_i) + l_2p_2(w_i|w_{i-1})+l_3p_3(w_i|w_{i-2},w_{i-1})$$\n",
    "pro $l_0+l_1+l_2+l_3 = 1$\n",
    "\n",
    "### Vypočítat cross-entropii\n",
    "Poupravit parametry vyhlazení následujícím způsobem: přidat $10\\%​$, $20\\%​$, $\\ldots​$,$90\\%​$, $95\\%​$ a $99\\%​$ rozdílu mezi trigram vyhlazení a $1.0​$ k jeho hodnotě, proporciálně uprav zbývající tři parametry (připomeňme, že mají součet 1!!). Potom nastavte trigram vyhlazovací parametr na $90\\%​$, $80\\%​$, $\\ldots​$, $10\\%​$, $0\\%​$ z jeho hodnoty a proporciálně uprav zbývající parametry. Vypočítejte cross-entropii na testovacích datech pro těchto 22 příkladů "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import random\n",
    "from enum import IntEnum\n",
    "import sys\n",
    "import math\n",
    "from tqdm import tqdm, tnrange, tqdm_notebook\n",
    "import time\n",
    "import pandas as pd\n",
    "from IPython.display import display, HTML\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Konstanty\n",
    "Nastavení hlavních parametrů"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "FILES = [\"TEXTEN1.txt\", \"TEXTCZ1.txt\"]\n",
    "TESTDATASIZE = 20*1000\n",
    "HELDOUTDATASIZE = 40*1000\n",
    "TRANSFORMTYPE = IntEnum('TRANSFORM', 'INCREASE DECREASE')\n",
    "TRANSFORMINC = [0,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,0.95,0.99]\n",
    "TRANSFORMDEC = [1,0.9,0.8,0.7,0.6,0.5,0.4,0.3,0.2,0.1,0]\n",
    "EPSILON = 1e-10\n",
    "DEBUG = False\n",
    "SEED = 84840014\n",
    "\n",
    "COUNTS_WORDS = {}\n",
    "COUNTS_PAIRS = {}\n",
    "COUNTS_TRIPLES = {}\n",
    "T = 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Načtení dat ze souboru\n",
    "Načtení dat do listu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def get_data(file):\n",
    "    with open(file, \"r\") as f:\n",
    "        data = np.array([x.replace(\"\\n\", \"\") for x in f.readlines()])\n",
    "        \n",
    "    if(DEBUG):\n",
    "        print(\"Metoda: get_data\")\n",
    "        print(\"Soubor: \",file)\n",
    "        print(\"Výsledek: \",data)\n",
    "        print()\n",
    "        \n",
    "    return data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Vytvoření datasetů a četností uni-,bi-,tri-gramů\n",
    "Vytvoření datasetů rozřezáním původního listu.\n",
    "Spočítání četností a uložení do slovníků typu (i,j,k):c(i,j,k)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def create_datasets(data):\n",
    "    test_data = data[-TESTDATASIZE:]\n",
    "    heldout_data = data[-(TESTDATASIZE+HELDOUTDATASIZE):-TESTDATASIZE]\n",
    "    train_data = data[:-(TESTDATASIZE+HELDOUTDATASIZE)]\n",
    "        \n",
    "    return (train_data,heldout_data,test_data)\n",
    "\n",
    "def create_counts(data):\n",
    "    # Odečíst 1 od četnosti posledního slova, protože to už nemá dvojici a já chci z toho počítat psti\n",
    "    words, counts = np.unique(data, return_counts=True)\n",
    "    counts_words = dict(zip(words,counts))\n",
    "    \n",
    "    counts_words[data[-1]] -= 1\n",
    "    \n",
    "    counts_pairs = dict()\n",
    "    for i in range(data.size - 1):\n",
    "        if (data[i],data[i+1]) in counts_pairs:\n",
    "            counts_pairs[(data[i],data[i+1])] += 1\n",
    "        else:\n",
    "            counts_pairs[(data[i],data[i+1])] = 1\n",
    "    \n",
    "    counts_triples = dict()\n",
    "    for i in range(data.size - 2):\n",
    "        if (data[i],data[i+1],data[i+2]) in counts_triples:\n",
    "            counts_triples[(data[i],data[i+1],data[i+2])] += 1\n",
    "        else:\n",
    "            counts_triples[(data[i],data[i+1],data[i+2])] = 1\n",
    "        \n",
    "    return (counts_words,counts_pairs,counts_triples)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pomocné funkce"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Výpočet pravděpodobností uni, bi, tri gramů\n",
    "Kontrola a řešení problémů s dělení 0 a následné vypočítání pravděpodobnosti podle vzorců. Při případném dělení 0 se používá pravděpodobnost $p_0$ viz slajdy předmětu."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def p0():\n",
    "    return 1/len(COUNTS_WORDS)\n",
    "\n",
    "def p1(w1):\n",
    "    if w1 not in COUNTS_WORDS:\n",
    "        return 0\n",
    "    return COUNTS_WORDS[w1]/T\n",
    "\n",
    "def p2(w1,w2):\n",
    "    if w1 not in COUNTS_WORDS or COUNTS_WORDS[w1]==0:\n",
    "        return p0()\n",
    "    if not (w1,w2) in COUNTS_PAIRS:\n",
    "        return 0\n",
    "    return COUNTS_PAIRS[(w1,w2)]/COUNTS_WORDS[w1]\n",
    "\n",
    "def p3(w1,w2,w3):\n",
    "    if (w1,w2) not in COUNTS_PAIRS or COUNTS_PAIRS[(w1,w2)] == 0:\n",
    "        return p0()\n",
    "    if not (w1,w2,w3) in COUNTS_TRIPLES:\n",
    "        return 0\n",
    "    return COUNTS_TRIPLES[(w1,w2,w3)]/COUNTS_PAIRS[(w1,w2)]\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Výpočet $p'_{\\lambda}(w_i|h_i)$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def p_lambda(w1,w2,w3,lambdas):\n",
    "    return lambdas[3]*p3(w1,w2,w3) \\\n",
    "          +lambdas[2]*p2(w2,w3) \\\n",
    "          +lambdas[1]*p1(w3) \\\n",
    "          +lambdas[0]*p0()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EM algoritmus\n",
    "Dokud se mění parametry o více než $\\epsilon$ provádí se výpočet, kde jsou spočítané očekávané hodnoty za použití aktuálních parametrů a následné jejich updatování."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def EM(epsilon, heldout_data):\n",
    "    lambdas = [0.25]*4\n",
    "    history = [lambdas]\n",
    "    \n",
    "    last_value = 0\n",
    "    with tqdm_notebook(total=1000,desc=\"EM\", leave=False) as pbar:\n",
    "    \n",
    "        while(True):\n",
    "            expected_counts = [0]*4\n",
    "\n",
    "            for i in range(heldout_data.size-2):\n",
    "                p_lam = p_lambda(heldout_data[i],heldout_data[i+1],heldout_data[i+2],lambdas)\n",
    "                if(p_lam==0):\n",
    "                    continue\n",
    "                expected_counts[0] += lambdas[0]*p0()\\\n",
    "                                     /p_lam\n",
    "                expected_counts[1] += lambdas[1]*p1(heldout_data[i+2])\\\n",
    "                                     /p_lam\n",
    "                expected_counts[2] += lambdas[2]*p2(heldout_data[i+1],heldout_data[i+2])\\\n",
    "                                     /p_lam\n",
    "                expected_counts[3] += lambdas[3]*p3(heldout_data[i],heldout_data[i+1],heldout_data[i+2])\\\n",
    "                                     /p_lam\n",
    "\n",
    "            new_lambdas = [x/sum(expected_counts) for x in expected_counts]\n",
    "\n",
    "            is_end = True\n",
    "            maximum = 0\n",
    "            for i in range(len(lambdas)):\n",
    "                if(abs(lambdas[i]-new_lambdas[i]) > maximum):\n",
    "                    maximum = abs(lambdas[i]-new_lambdas[i])\n",
    "                if(abs(lambdas[i]-new_lambdas[i])>epsilon):\n",
    "                    is_end = False\n",
    "                    break\n",
    "\n",
    "            if is_end:\n",
    "                pbar.update(1000-pbar.last_print_n)\n",
    "                break\n",
    "\n",
    "            lambdas = new_lambdas\n",
    "            history.append(lambdas)\n",
    "            \n",
    "            #Progress bar\n",
    "            new_value = 1/math.log(EPSILON,10)*math.log(maximum,10)\n",
    "            pbar.update(int((new_value - last_value)*1000) if new_value - last_value > 0 else 0)\n",
    "            last_value = new_value if new_value > last_value else last_value\n",
    "    \n",
    "        \n",
    "    \n",
    "    return (lambdas,history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Cross-Entropie"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def entropy(data, lambdas):\n",
    "    H = 0\n",
    "    for i in range(data.size-2):\n",
    "        H -= math.log2(p_lambda(data[i],data[i+1],data[i+2],lambdas))\n",
    "        \n",
    "    return H/(data.size-2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Změna lambdy\n",
    "Pomocné metody k změnění parametrů o určitá procenta. Vypočte se nová $\\lambda_3$ a poté se poměrově zbytek hodnoty přiřadí ostatním parametrům."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def lambda_increase(lambdas, perc):\n",
    "    lambdas[-1] += (1 - lambdas[-1])*perc\n",
    "    ratio = (sum(lambdas)-lambdas[-1])/(1-lambdas[-1])\n",
    "    lambdas = [x/ratio for x in lambdas]\n",
    "    lambdas[-1] *= ratio\n",
    "    return lambdas\n",
    "\n",
    "def lambda_decrease(lambdas, perc):\n",
    "    lambdas[-1] *= perc\n",
    "    ratio = (sum(lambdas)-lambdas[-1])/(1-lambdas[-1])\n",
    "    lambdas = [x/ratio for x in lambdas]\n",
    "    lambdas[-1] *= ratio\n",
    "    return lambdas\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Zobrazení výsledků"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def show_results(results):\n",
    "    for i in range(len(FILES)):\n",
    "        for j in range(len(TRANSFORMTYPE)):\n",
    "            if (j+1)==int(TRANSFORMTYPE.INCREASE):\n",
    "                df1 = pd.DataFrame(results[i,j,:len(TRANSFORMINC),:],index=TRANSFORMINC, columns=[\"$\\lambda_0$\",\"$\\lambda_1$\",\"$\\lambda_2$\",\"$\\lambda_3$\",\"Entropie\"])\n",
    "                graph = pd.DataFrame(results[i,j,:len(TRANSFORMINC),-1],index=TRANSFORMINC, columns=[str(FILES[i] + \" - \" + str(TRANSFORMTYPE(j+1)).split(\".\")[1])])\n",
    "            if (j+1)==int(TRANSFORMTYPE.DECREASE):\n",
    "                df1 = pd.DataFrame(results[i,j,:len(TRANSFORMDEC),:],index=TRANSFORMDEC, columns=[\"$\\lambda_0$\",\"$\\lambda_1$\",\"$\\lambda_2$\",\"$\\lambda_3$\",\"Entropie\"])\n",
    "                graph = pd.DataFrame(results[i,j,:len(TRANSFORMDEC),-1],index=[1-x for x in TRANSFORMDEC], columns=[str(FILES[i] + \" - \" + str(TRANSFORMTYPE(j+1)).split(\".\")[1])])\n",
    "                \n",
    "            plt.plot(graph, label=str(FILES[i] + \" - \" + str(TRANSFORMTYPE(j+1)).split(\".\")[1]))\n",
    "            plt.plot(graph, 'k|')\n",
    "            \n",
    "            for col in df1.columns:\n",
    "                df1[col] = df1[col].map('{:,.10f}'.format)\n",
    "            \n",
    "            display(HTML(\"<h2>\" + FILES[i] + \" - \" + str(TRANSFORMTYPE(j+1)).split(\".\")[1] + \"</h2>\"))\n",
    "            display(df1)\n",
    "            \n",
    "    plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)\n",
    "    plt.xlabel('Zvýšení/snížení $\\lambda_3$')\n",
    "    plt.ylabel('Cross-Entropie')\n",
    "    plt.title('Graf Cross-Entropie')\n",
    "    plt.show()\n",
    "\n",
    "def show_lambdas(lambdas):    \n",
    "    for i in range(len(FILES)):\n",
    "        df1 = pd.DataFrame(lambdas[i], columns=[\"$\\lambda_0$\",\"$\\lambda_1$\",\"$\\lambda_2$\",\"$\\lambda_3$\"])\n",
    "        for col in df1.columns:\n",
    "            df1[col] = df1[col].map('{:,.10f}'.format)\n",
    "        display(HTML(\"<h2>\" + FILES[i] + \"</h2>\"))\n",
    "        display(df1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Graf pokrytí\n",
    "Výpočet a vykreslení grafu. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def coverage(test_data):\n",
    "    test_counts = create_counts(test_data)\n",
    "    train_counts = (COUNTS_WORDS,COUNTS_PAIRS,COUNTS_TRIPLES)\n",
    "    cov = [0]*3\n",
    "    for i in range(len(cov)):\n",
    "        for t in test_counts[i]:\n",
    "            if t in train_counts[i]:\n",
    "                cov[i] += test_counts[i][t]\n",
    "        cov[i] /= sum([test_counts[i][x] for x in test_counts[i]])\n",
    "    return cov\n",
    "\n",
    "def show_coverage(results):\n",
    "    df1 = pd.DataFrame(results, columns=[\"Uni-gram\", \"Bi-gram\", \"Tri-gram\"], index=FILES)\n",
    "    df1 = df1.transpose()\n",
    "    display(df1)\n",
    "    \n",
    "    colors = [\"r\",\"b\",\"g\"]\n",
    "    n_groups = 3\n",
    "\n",
    "    index = np.arange(n_groups)\n",
    "    bar_width = 0.8/len(FILES)\n",
    "\n",
    "    opacity = 0.4\n",
    "    error_config = {'ecolor': '0.3'}\n",
    "    \n",
    "    for i in range(len(results)):\n",
    "        plt.bar(index + i*bar_width, results[i], bar_width,\n",
    "                         alpha=opacity,\n",
    "                         color=colors[i%len(colors)],\n",
    "                         label=FILES[i])\n",
    "\n",
    "    plt.xlabel('n-gram')\n",
    "    plt.ylabel('Pokrytí')\n",
    "    plt.title('Graf pokrytí')\n",
    "    plt.xticks(index + bar_width*len(FILES) / 2, ('Uni-gram', 'Bi-gram', 'Tri-gram'))\n",
    "    plt.legend()\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Main\n",
    "### Hlavní cyklus\n",
    "Projdou se všechny soubory.\n",
    "Z každého souboru jsou vytvořeny datasety, spočítány uni-,bi-,tri-gramy a je spuštěn EM algoritmus.\n",
    "Poté se spočítána cross-entropie pro všechna požadovaná nastavení parametrů. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Time:  3 m  42 s\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<h1>Průběžné výsledky EM algoritmu</h1>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h2>TEXTEN1.txt</h2>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$\\lambda_0$</th>\n",
       "      <th>$\\lambda_1$</th>\n",
       "      <th>$\\lambda_2$</th>\n",
       "      <th>$\\lambda_3$</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.1167225989</td>\n",
       "      <td>0.2514594534</td>\n",
       "      <td>0.3922685188</td>\n",
       "      <td>0.2395494289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0847611305</td>\n",
       "      <td>0.2490138807</td>\n",
       "      <td>0.4500333134</td>\n",
       "      <td>0.2161916753</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0751590871</td>\n",
       "      <td>0.2503157295</td>\n",
       "      <td>0.4729723925</td>\n",
       "      <td>0.2015527909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0718728325</td>\n",
       "      <td>0.2519520613</td>\n",
       "      <td>0.4827689266</td>\n",
       "      <td>0.1934061796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.0706822968</td>\n",
       "      <td>0.2530183923</td>\n",
       "      <td>0.4873265040</td>\n",
       "      <td>0.1889728068</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.0702485407</td>\n",
       "      <td>0.2535824570</td>\n",
       "      <td>0.4896026037</td>\n",
       "      <td>0.1865663986</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.0700982059</td>\n",
       "      <td>0.2538479539</td>\n",
       "      <td>0.4907970868</td>\n",
       "      <td>0.1852567535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.0700537346</td>\n",
       "      <td>0.2539612547</td>\n",
       "      <td>0.4914440006</td>\n",
       "      <td>0.1845410101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.0700468531</td>\n",
       "      <td>0.2540039738</td>\n",
       "      <td>0.4918011264</td>\n",
       "      <td>0.1841480466</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.0700515078</td>\n",
       "      <td>0.2540166137</td>\n",
       "      <td>0.4920005648</td>\n",
       "      <td>0.1839313136</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.0700582425</td>\n",
       "      <td>0.2540177622</td>\n",
       "      <td>0.4921127374</td>\n",
       "      <td>0.1838112579</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.0700640547</td>\n",
       "      <td>0.2540153424</td>\n",
       "      <td>0.4921761174</td>\n",
       "      <td>0.1837444855</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.0700683322</td>\n",
       "      <td>0.2540124177</td>\n",
       "      <td>0.4922120399</td>\n",
       "      <td>0.1837072102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.0700712452</td>\n",
       "      <td>0.2540099780</td>\n",
       "      <td>0.4922324455</td>\n",
       "      <td>0.1836863313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.0700731406</td>\n",
       "      <td>0.2540082019</td>\n",
       "      <td>0.4922440565</td>\n",
       "      <td>0.1836746010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.0700743378</td>\n",
       "      <td>0.2540069974</td>\n",
       "      <td>0.4922506721</td>\n",
       "      <td>0.1836679927</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.0700750784</td>\n",
       "      <td>0.2540062151</td>\n",
       "      <td>0.4922544455</td>\n",
       "      <td>0.1836642610</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.0700755298</td>\n",
       "      <td>0.2540057213</td>\n",
       "      <td>0.4922565997</td>\n",
       "      <td>0.1836621492</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.0700758017</td>\n",
       "      <td>0.2540054159</td>\n",
       "      <td>0.4922578305</td>\n",
       "      <td>0.1836609519</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.0700759641</td>\n",
       "      <td>0.2540052298</td>\n",
       "      <td>0.4922585341</td>\n",
       "      <td>0.1836602720</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.0700760604</td>\n",
       "      <td>0.2540051177</td>\n",
       "      <td>0.4922589366</td>\n",
       "      <td>0.1836598853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.0700761172</td>\n",
       "      <td>0.2540050507</td>\n",
       "      <td>0.4922591670</td>\n",
       "      <td>0.1836596651</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.0700761505</td>\n",
       "      <td>0.2540050110</td>\n",
       "      <td>0.4922592988</td>\n",
       "      <td>0.1836595396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.0700761700</td>\n",
       "      <td>0.2540049876</td>\n",
       "      <td>0.4922593744</td>\n",
       "      <td>0.1836594680</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.0700761814</td>\n",
       "      <td>0.2540049739</td>\n",
       "      <td>0.4922594176</td>\n",
       "      <td>0.1836594271</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.0700761880</td>\n",
       "      <td>0.2540049658</td>\n",
       "      <td>0.4922594424</td>\n",
       "      <td>0.1836594037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.0700761918</td>\n",
       "      <td>0.2540049611</td>\n",
       "      <td>0.4922594566</td>\n",
       "      <td>0.1836593904</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.0700761941</td>\n",
       "      <td>0.2540049584</td>\n",
       "      <td>0.4922594648</td>\n",
       "      <td>0.1836593827</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.0700761954</td>\n",
       "      <td>0.2540049568</td>\n",
       "      <td>0.4922594694</td>\n",
       "      <td>0.1836593784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.0700761961</td>\n",
       "      <td>0.2540049559</td>\n",
       "      <td>0.4922594721</td>\n",
       "      <td>0.1836593759</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.0700761965</td>\n",
       "      <td>0.2540049554</td>\n",
       "      <td>0.4922594737</td>\n",
       "      <td>0.1836593744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.0700761968</td>\n",
       "      <td>0.2540049551</td>\n",
       "      <td>0.4922594745</td>\n",
       "      <td>0.1836593736</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.0700761969</td>\n",
       "      <td>0.2540049549</td>\n",
       "      <td>0.4922594750</td>\n",
       "      <td>0.1836593731</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.0700761970</td>\n",
       "      <td>0.2540049548</td>\n",
       "      <td>0.4922594753</td>\n",
       "      <td>0.1836593729</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.0700761971</td>\n",
       "      <td>0.2540049547</td>\n",
       "      <td>0.4922594755</td>\n",
       "      <td>0.1836593727</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     $\\lambda_0$   $\\lambda_1$   $\\lambda_2$   $\\lambda_3$\n",
       "0   0.2500000000  0.2500000000  0.2500000000  0.2500000000\n",
       "1   0.1167225989  0.2514594534  0.3922685188  0.2395494289\n",
       "2   0.0847611305  0.2490138807  0.4500333134  0.2161916753\n",
       "3   0.0751590871  0.2503157295  0.4729723925  0.2015527909\n",
       "4   0.0718728325  0.2519520613  0.4827689266  0.1934061796\n",
       "5   0.0706822968  0.2530183923  0.4873265040  0.1889728068\n",
       "6   0.0702485407  0.2535824570  0.4896026037  0.1865663986\n",
       "7   0.0700982059  0.2538479539  0.4907970868  0.1852567535\n",
       "8   0.0700537346  0.2539612547  0.4914440006  0.1845410101\n",
       "9   0.0700468531  0.2540039738  0.4918011264  0.1841480466\n",
       "10  0.0700515078  0.2540166137  0.4920005648  0.1839313136\n",
       "11  0.0700582425  0.2540177622  0.4921127374  0.1838112579\n",
       "12  0.0700640547  0.2540153424  0.4921761174  0.1837444855\n",
       "13  0.0700683322  0.2540124177  0.4922120399  0.1837072102\n",
       "14  0.0700712452  0.2540099780  0.4922324455  0.1836863313\n",
       "15  0.0700731406  0.2540082019  0.4922440565  0.1836746010\n",
       "16  0.0700743378  0.2540069974  0.4922506721  0.1836679927\n",
       "17  0.0700750784  0.2540062151  0.4922544455  0.1836642610\n",
       "18  0.0700755298  0.2540057213  0.4922565997  0.1836621492\n",
       "19  0.0700758017  0.2540054159  0.4922578305  0.1836609519\n",
       "20  0.0700759641  0.2540052298  0.4922585341  0.1836602720\n",
       "21  0.0700760604  0.2540051177  0.4922589366  0.1836598853\n",
       "22  0.0700761172  0.2540050507  0.4922591670  0.1836596651\n",
       "23  0.0700761505  0.2540050110  0.4922592988  0.1836595396\n",
       "24  0.0700761700  0.2540049876  0.4922593744  0.1836594680\n",
       "25  0.0700761814  0.2540049739  0.4922594176  0.1836594271\n",
       "26  0.0700761880  0.2540049658  0.4922594424  0.1836594037\n",
       "27  0.0700761918  0.2540049611  0.4922594566  0.1836593904\n",
       "28  0.0700761941  0.2540049584  0.4922594648  0.1836593827\n",
       "29  0.0700761954  0.2540049568  0.4922594694  0.1836593784\n",
       "30  0.0700761961  0.2540049559  0.4922594721  0.1836593759\n",
       "31  0.0700761965  0.2540049554  0.4922594737  0.1836593744\n",
       "32  0.0700761968  0.2540049551  0.4922594745  0.1836593736\n",
       "33  0.0700761969  0.2540049549  0.4922594750  0.1836593731\n",
       "34  0.0700761970  0.2540049548  0.4922594753  0.1836593729\n",
       "35  0.0700761971  0.2540049547  0.4922594755  0.1836593727"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h2>TEXTCZ1.txt</h2>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$\\lambda_0$</th>\n",
       "      <th>$\\lambda_1$</th>\n",
       "      <th>$\\lambda_2$</th>\n",
       "      <th>$\\lambda_3$</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.1899333114</td>\n",
       "      <td>0.3565384785</td>\n",
       "      <td>0.2513176208</td>\n",
       "      <td>0.2022105893</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.1641245035</td>\n",
       "      <td>0.4007131077</td>\n",
       "      <td>0.2486312278</td>\n",
       "      <td>0.1865311609</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.1526923559</td>\n",
       "      <td>0.4177170910</td>\n",
       "      <td>0.2471064413</td>\n",
       "      <td>0.1824841118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.1472225773</td>\n",
       "      <td>0.4242560715</td>\n",
       "      <td>0.2462761313</td>\n",
       "      <td>0.1822452199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.1443831160</td>\n",
       "      <td>0.4268411933</td>\n",
       "      <td>0.2457541713</td>\n",
       "      <td>0.1830215193</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.1427990662</td>\n",
       "      <td>0.4279156949</td>\n",
       "      <td>0.2453963927</td>\n",
       "      <td>0.1838888463</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.1418642966</td>\n",
       "      <td>0.4283951015</td>\n",
       "      <td>0.2451463654</td>\n",
       "      <td>0.1845942364</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.1412903092</td>\n",
       "      <td>0.4286279174</td>\n",
       "      <td>0.2449731459</td>\n",
       "      <td>0.1851086276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.1409285256</td>\n",
       "      <td>0.4287510766</td>\n",
       "      <td>0.2448548613</td>\n",
       "      <td>0.1854655366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.1406967386</td>\n",
       "      <td>0.4288211678</td>\n",
       "      <td>0.2447751108</td>\n",
       "      <td>0.1857069828</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.1405467638</td>\n",
       "      <td>0.4288632757</td>\n",
       "      <td>0.2447218471</td>\n",
       "      <td>0.1858681135</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.1404491556</td>\n",
       "      <td>0.4288894984</td>\n",
       "      <td>0.2446865041</td>\n",
       "      <td>0.1859748419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.1403854121</td>\n",
       "      <td>0.4289061949</td>\n",
       "      <td>0.2446631528</td>\n",
       "      <td>0.1860452402</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.1403437019</td>\n",
       "      <td>0.4289169656</td>\n",
       "      <td>0.2446477668</td>\n",
       "      <td>0.1860915657</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.1403163782</td>\n",
       "      <td>0.4289239657</td>\n",
       "      <td>0.2446376465</td>\n",
       "      <td>0.1861220096</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.1402984674</td>\n",
       "      <td>0.4289285343</td>\n",
       "      <td>0.2446309969</td>\n",
       "      <td>0.1861420013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.1402867225</td>\n",
       "      <td>0.4289315231</td>\n",
       "      <td>0.2446266305</td>\n",
       "      <td>0.1861551239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.1402790193</td>\n",
       "      <td>0.4289334809</td>\n",
       "      <td>0.2446237644</td>\n",
       "      <td>0.1861637355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.1402739663</td>\n",
       "      <td>0.4289347642</td>\n",
       "      <td>0.2446218836</td>\n",
       "      <td>0.1861693860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.1402706515</td>\n",
       "      <td>0.4289356057</td>\n",
       "      <td>0.2446206494</td>\n",
       "      <td>0.1861730933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.1402684770</td>\n",
       "      <td>0.4289361577</td>\n",
       "      <td>0.2446198397</td>\n",
       "      <td>0.1861755256</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.1402670504</td>\n",
       "      <td>0.4289365198</td>\n",
       "      <td>0.2446193085</td>\n",
       "      <td>0.1861771213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.1402661146</td>\n",
       "      <td>0.4289367573</td>\n",
       "      <td>0.2446189600</td>\n",
       "      <td>0.1861781682</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.1402655006</td>\n",
       "      <td>0.4289369131</td>\n",
       "      <td>0.2446187313</td>\n",
       "      <td>0.1861788550</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.1402650978</td>\n",
       "      <td>0.4289370153</td>\n",
       "      <td>0.2446185813</td>\n",
       "      <td>0.1861793055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.1402648336</td>\n",
       "      <td>0.4289370824</td>\n",
       "      <td>0.2446184829</td>\n",
       "      <td>0.1861796011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.1402646602</td>\n",
       "      <td>0.4289371264</td>\n",
       "      <td>0.2446184184</td>\n",
       "      <td>0.1861797951</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.1402645465</td>\n",
       "      <td>0.4289371552</td>\n",
       "      <td>0.2446183760</td>\n",
       "      <td>0.1861799223</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.1402644719</td>\n",
       "      <td>0.4289371742</td>\n",
       "      <td>0.2446183482</td>\n",
       "      <td>0.1861800057</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.1402644229</td>\n",
       "      <td>0.4289371866</td>\n",
       "      <td>0.2446183300</td>\n",
       "      <td>0.1861800605</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.1402643908</td>\n",
       "      <td>0.4289371948</td>\n",
       "      <td>0.2446183180</td>\n",
       "      <td>0.1861800964</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.1402643697</td>\n",
       "      <td>0.4289372001</td>\n",
       "      <td>0.2446183102</td>\n",
       "      <td>0.1861801200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.1402643559</td>\n",
       "      <td>0.4289372036</td>\n",
       "      <td>0.2446183050</td>\n",
       "      <td>0.1861801355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.1402643468</td>\n",
       "      <td>0.4289372059</td>\n",
       "      <td>0.2446183017</td>\n",
       "      <td>0.1861801456</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.1402643409</td>\n",
       "      <td>0.4289372074</td>\n",
       "      <td>0.2446182994</td>\n",
       "      <td>0.1861801522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>0.1402643370</td>\n",
       "      <td>0.4289372084</td>\n",
       "      <td>0.2446182980</td>\n",
       "      <td>0.1861801566</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>0.1402643344</td>\n",
       "      <td>0.4289372091</td>\n",
       "      <td>0.2446182970</td>\n",
       "      <td>0.1861801595</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>0.1402643327</td>\n",
       "      <td>0.4289372095</td>\n",
       "      <td>0.2446182964</td>\n",
       "      <td>0.1861801614</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>0.1402643316</td>\n",
       "      <td>0.4289372098</td>\n",
       "      <td>0.2446182960</td>\n",
       "      <td>0.1861801626</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>0.1402643309</td>\n",
       "      <td>0.4289372099</td>\n",
       "      <td>0.2446182957</td>\n",
       "      <td>0.1861801634</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>0.1402643304</td>\n",
       "      <td>0.4289372101</td>\n",
       "      <td>0.2446182956</td>\n",
       "      <td>0.1861801639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>0.1402643301</td>\n",
       "      <td>0.4289372101</td>\n",
       "      <td>0.2446182954</td>\n",
       "      <td>0.1861801643</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>0.1402643299</td>\n",
       "      <td>0.4289372102</td>\n",
       "      <td>0.2446182954</td>\n",
       "      <td>0.1861801645</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>0.1402643298</td>\n",
       "      <td>0.4289372102</td>\n",
       "      <td>0.2446182953</td>\n",
       "      <td>0.1861801647</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     $\\lambda_0$   $\\lambda_1$   $\\lambda_2$   $\\lambda_3$\n",
       "0   0.2500000000  0.2500000000  0.2500000000  0.2500000000\n",
       "1   0.1899333114  0.3565384785  0.2513176208  0.2022105893\n",
       "2   0.1641245035  0.4007131077  0.2486312278  0.1865311609\n",
       "3   0.1526923559  0.4177170910  0.2471064413  0.1824841118\n",
       "4   0.1472225773  0.4242560715  0.2462761313  0.1822452199\n",
       "5   0.1443831160  0.4268411933  0.2457541713  0.1830215193\n",
       "6   0.1427990662  0.4279156949  0.2453963927  0.1838888463\n",
       "7   0.1418642966  0.4283951015  0.2451463654  0.1845942364\n",
       "8   0.1412903092  0.4286279174  0.2449731459  0.1851086276\n",
       "9   0.1409285256  0.4287510766  0.2448548613  0.1854655366\n",
       "10  0.1406967386  0.4288211678  0.2447751108  0.1857069828\n",
       "11  0.1405467638  0.4288632757  0.2447218471  0.1858681135\n",
       "12  0.1404491556  0.4288894984  0.2446865041  0.1859748419\n",
       "13  0.1403854121  0.4289061949  0.2446631528  0.1860452402\n",
       "14  0.1403437019  0.4289169656  0.2446477668  0.1860915657\n",
       "15  0.1403163782  0.4289239657  0.2446376465  0.1861220096\n",
       "16  0.1402984674  0.4289285343  0.2446309969  0.1861420013\n",
       "17  0.1402867225  0.4289315231  0.2446266305  0.1861551239\n",
       "18  0.1402790193  0.4289334809  0.2446237644  0.1861637355\n",
       "19  0.1402739663  0.4289347642  0.2446218836  0.1861693860\n",
       "20  0.1402706515  0.4289356057  0.2446206494  0.1861730933\n",
       "21  0.1402684770  0.4289361577  0.2446198397  0.1861755256\n",
       "22  0.1402670504  0.4289365198  0.2446193085  0.1861771213\n",
       "23  0.1402661146  0.4289367573  0.2446189600  0.1861781682\n",
       "24  0.1402655006  0.4289369131  0.2446187313  0.1861788550\n",
       "25  0.1402650978  0.4289370153  0.2446185813  0.1861793055\n",
       "26  0.1402648336  0.4289370824  0.2446184829  0.1861796011\n",
       "27  0.1402646602  0.4289371264  0.2446184184  0.1861797951\n",
       "28  0.1402645465  0.4289371552  0.2446183760  0.1861799223\n",
       "29  0.1402644719  0.4289371742  0.2446183482  0.1861800057\n",
       "30  0.1402644229  0.4289371866  0.2446183300  0.1861800605\n",
       "31  0.1402643908  0.4289371948  0.2446183180  0.1861800964\n",
       "32  0.1402643697  0.4289372001  0.2446183102  0.1861801200\n",
       "33  0.1402643559  0.4289372036  0.2446183050  0.1861801355\n",
       "34  0.1402643468  0.4289372059  0.2446183017  0.1861801456\n",
       "35  0.1402643409  0.4289372074  0.2446182994  0.1861801522\n",
       "36  0.1402643370  0.4289372084  0.2446182980  0.1861801566\n",
       "37  0.1402643344  0.4289372091  0.2446182970  0.1861801595\n",
       "38  0.1402643327  0.4289372095  0.2446182964  0.1861801614\n",
       "39  0.1402643316  0.4289372098  0.2446182960  0.1861801626\n",
       "40  0.1402643309  0.4289372099  0.2446182957  0.1861801634\n",
       "41  0.1402643304  0.4289372101  0.2446182956  0.1861801639\n",
       "42  0.1402643301  0.4289372101  0.2446182954  0.1861801643\n",
       "43  0.1402643299  0.4289372102  0.2446182954  0.1861801645\n",
       "44  0.1402643298  0.4289372102  0.2446182953  0.1861801647"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h1>Průběžné výsledky EM algoritmu na trénovacích datech</h1>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h2>TEXTEN1.txt</h2>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$\\lambda_0$</th>\n",
       "      <th>$\\lambda_1$</th>\n",
       "      <th>$\\lambda_2$</th>\n",
       "      <th>$\\lambda_3$</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0026777600</td>\n",
       "      <td>0.0348104979</td>\n",
       "      <td>0.2006972440</td>\n",
       "      <td>0.7618144981</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0000131842</td>\n",
       "      <td>0.0026745507</td>\n",
       "      <td>0.0783927207</td>\n",
       "      <td>0.9189195444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0000000585</td>\n",
       "      <td>0.0001956542</td>\n",
       "      <td>0.0287546342</td>\n",
       "      <td>0.9710496531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0000000003</td>\n",
       "      <td>0.0000141677</td>\n",
       "      <td>0.0104382616</td>\n",
       "      <td>0.9895475704</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000010231</td>\n",
       "      <td>0.0037802144</td>\n",
       "      <td>0.9962187624</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000738</td>\n",
       "      <td>0.0013680878</td>\n",
       "      <td>0.9986318384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000053</td>\n",
       "      <td>0.0004950145</td>\n",
       "      <td>0.9995049802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000004</td>\n",
       "      <td>0.0001790975</td>\n",
       "      <td>0.9998209021</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000647962</td>\n",
       "      <td>0.9999352037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000234426</td>\n",
       "      <td>0.9999765574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000084813</td>\n",
       "      <td>0.9999915187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000030684</td>\n",
       "      <td>0.9999969316</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000011101</td>\n",
       "      <td>0.9999988899</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000004016</td>\n",
       "      <td>0.9999995984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000001453</td>\n",
       "      <td>0.9999998547</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000526</td>\n",
       "      <td>0.9999999474</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000190</td>\n",
       "      <td>0.9999999810</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000069</td>\n",
       "      <td>0.9999999931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000025</td>\n",
       "      <td>0.9999999975</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000009</td>\n",
       "      <td>0.9999999991</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000003</td>\n",
       "      <td>0.9999999997</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000001</td>\n",
       "      <td>0.9999999999</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     $\\lambda_0$   $\\lambda_1$   $\\lambda_2$   $\\lambda_3$\n",
       "0   0.2500000000  0.2500000000  0.2500000000  0.2500000000\n",
       "1   0.0026777600  0.0348104979  0.2006972440  0.7618144981\n",
       "2   0.0000131842  0.0026745507  0.0783927207  0.9189195444\n",
       "3   0.0000000585  0.0001956542  0.0287546342  0.9710496531\n",
       "4   0.0000000003  0.0000141677  0.0104382616  0.9895475704\n",
       "5   0.0000000000  0.0000010231  0.0037802144  0.9962187624\n",
       "6   0.0000000000  0.0000000738  0.0013680878  0.9986318384\n",
       "7   0.0000000000  0.0000000053  0.0004950145  0.9995049802\n",
       "8   0.0000000000  0.0000000004  0.0001790975  0.9998209021\n",
       "9   0.0000000000  0.0000000000  0.0000647962  0.9999352037\n",
       "10  0.0000000000  0.0000000000  0.0000234426  0.9999765574\n",
       "11  0.0000000000  0.0000000000  0.0000084813  0.9999915187\n",
       "12  0.0000000000  0.0000000000  0.0000030684  0.9999969316\n",
       "13  0.0000000000  0.0000000000  0.0000011101  0.9999988899\n",
       "14  0.0000000000  0.0000000000  0.0000004016  0.9999995984\n",
       "15  0.0000000000  0.0000000000  0.0000001453  0.9999998547\n",
       "16  0.0000000000  0.0000000000  0.0000000526  0.9999999474\n",
       "17  0.0000000000  0.0000000000  0.0000000190  0.9999999810\n",
       "18  0.0000000000  0.0000000000  0.0000000069  0.9999999931\n",
       "19  0.0000000000  0.0000000000  0.0000000025  0.9999999975\n",
       "20  0.0000000000  0.0000000000  0.0000000009  0.9999999991\n",
       "21  0.0000000000  0.0000000000  0.0000000003  0.9999999997\n",
       "22  0.0000000000  0.0000000000  0.0000000001  0.9999999999"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h2>TEXTCZ1.txt</h2>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$\\lambda_0$</th>\n",
       "      <th>$\\lambda_1$</th>\n",
       "      <th>$\\lambda_2$</th>\n",
       "      <th>$\\lambda_3$</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "      <td>0.2500000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0002187048</td>\n",
       "      <td>0.0118670333</td>\n",
       "      <td>0.2053978789</td>\n",
       "      <td>0.7825163829</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0000000818</td>\n",
       "      <td>0.0002775057</td>\n",
       "      <td>0.0786837976</td>\n",
       "      <td>0.9210386149</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000062118</td>\n",
       "      <td>0.0289201173</td>\n",
       "      <td>0.9710736708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000001379</td>\n",
       "      <td>0.0105331000</td>\n",
       "      <td>0.9894667621</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000031</td>\n",
       "      <td>0.0038265598</td>\n",
       "      <td>0.9961734371</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000001</td>\n",
       "      <td>0.0013891092</td>\n",
       "      <td>0.9986108908</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0005041512</td>\n",
       "      <td>0.9994958488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0001829573</td>\n",
       "      <td>0.9998170427</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000663936</td>\n",
       "      <td>0.9999336064</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000240934</td>\n",
       "      <td>0.9999759066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000087432</td>\n",
       "      <td>0.9999912568</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000031728</td>\n",
       "      <td>0.9999968272</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000011514</td>\n",
       "      <td>0.9999988486</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000004178</td>\n",
       "      <td>0.9999995822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000001516</td>\n",
       "      <td>0.9999998484</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000550</td>\n",
       "      <td>0.9999999450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000200</td>\n",
       "      <td>0.9999999800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000072</td>\n",
       "      <td>0.9999999928</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000026</td>\n",
       "      <td>0.9999999974</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000010</td>\n",
       "      <td>0.9999999990</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000003</td>\n",
       "      <td>0.9999999997</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>0.0000000001</td>\n",
       "      <td>0.9999999999</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     $\\lambda_0$   $\\lambda_1$   $\\lambda_2$   $\\lambda_3$\n",
       "0   0.2500000000  0.2500000000  0.2500000000  0.2500000000\n",
       "1   0.0002187048  0.0118670333  0.2053978789  0.7825163829\n",
       "2   0.0000000818  0.0002775057  0.0786837976  0.9210386149\n",
       "3   0.0000000000  0.0000062118  0.0289201173  0.9710736708\n",
       "4   0.0000000000  0.0000001379  0.0105331000  0.9894667621\n",
       "5   0.0000000000  0.0000000031  0.0038265598  0.9961734371\n",
       "6   0.0000000000  0.0000000001  0.0013891092  0.9986108908\n",
       "7   0.0000000000  0.0000000000  0.0005041512  0.9994958488\n",
       "8   0.0000000000  0.0000000000  0.0001829573  0.9998170427\n",
       "9   0.0000000000  0.0000000000  0.0000663936  0.9999336064\n",
       "10  0.0000000000  0.0000000000  0.0000240934  0.9999759066\n",
       "11  0.0000000000  0.0000000000  0.0000087432  0.9999912568\n",
       "12  0.0000000000  0.0000000000  0.0000031728  0.9999968272\n",
       "13  0.0000000000  0.0000000000  0.0000011514  0.9999988486\n",
       "14  0.0000000000  0.0000000000  0.0000004178  0.9999995822\n",
       "15  0.0000000000  0.0000000000  0.0000001516  0.9999998484\n",
       "16  0.0000000000  0.0000000000  0.0000000550  0.9999999450\n",
       "17  0.0000000000  0.0000000000  0.0000000200  0.9999999800\n",
       "18  0.0000000000  0.0000000000  0.0000000072  0.9999999928\n",
       "19  0.0000000000  0.0000000000  0.0000000026  0.9999999974\n",
       "20  0.0000000000  0.0000000000  0.0000000010  0.9999999990\n",
       "21  0.0000000000  0.0000000000  0.0000000003  0.9999999997\n",
       "22  0.0000000000  0.0000000000  0.0000000001  0.9999999999"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h1>Nastavení $\\lambda_i$ a entropie</h1>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h2>TEXTEN1.txt - INCREASE</h2>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$\\lambda_0$</th>\n",
       "      <th>$\\lambda_1$</th>\n",
       "      <th>$\\lambda_2$</th>\n",
       "      <th>$\\lambda_3$</th>\n",
       "      <th>Entropie</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.00</th>\n",
       "      <td>0.0700761971</td>\n",
       "      <td>0.2540049547</td>\n",
       "      <td>0.4922594755</td>\n",
       "      <td>0.1836593727</td>\n",
       "      <td>7.4685164076</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.10</th>\n",
       "      <td>0.0630685773</td>\n",
       "      <td>0.2286044593</td>\n",
       "      <td>0.4430335280</td>\n",
       "      <td>0.2652934354</td>\n",
       "      <td>7.4700973606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.20</th>\n",
       "      <td>0.0560609576</td>\n",
       "      <td>0.2032039638</td>\n",
       "      <td>0.3938075804</td>\n",
       "      <td>0.3469274982</td>\n",
       "      <td>7.4969927508</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.30</th>\n",
       "      <td>0.0490533379</td>\n",
       "      <td>0.1778034683</td>\n",
       "      <td>0.3445816329</td>\n",
       "      <td>0.4285615609</td>\n",
       "      <td>7.5467698528</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.40</th>\n",
       "      <td>0.0420457182</td>\n",
       "      <td>0.1524029728</td>\n",
       "      <td>0.2953556853</td>\n",
       "      <td>0.5101956236</td>\n",
       "      <td>7.6204790432</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.50</th>\n",
       "      <td>0.0350380985</td>\n",
       "      <td>0.1270024774</td>\n",
       "      <td>0.2461297378</td>\n",
       "      <td>0.5918296864</td>\n",
       "      <td>7.7223233454</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.60</th>\n",
       "      <td>0.0280304788</td>\n",
       "      <td>0.1016019819</td>\n",
       "      <td>0.1969037902</td>\n",
       "      <td>0.6734637491</td>\n",
       "      <td>7.8610268679</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.70</th>\n",
       "      <td>0.0210228591</td>\n",
       "      <td>0.0762014864</td>\n",
       "      <td>0.1476778427</td>\n",
       "      <td>0.7550978118</td>\n",
       "      <td>8.0541442064</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.80</th>\n",
       "      <td>0.0140152394</td>\n",
       "      <td>0.0508009909</td>\n",
       "      <td>0.0984518951</td>\n",
       "      <td>0.8367318745</td>\n",
       "      <td>8.3418290591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.90</th>\n",
       "      <td>0.0070076197</td>\n",
       "      <td>0.0254004955</td>\n",
       "      <td>0.0492259476</td>\n",
       "      <td>0.9183659373</td>\n",
       "      <td>8.8510590988</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.95</th>\n",
       "      <td>0.0035038099</td>\n",
       "      <td>0.0127002477</td>\n",
       "      <td>0.0246129738</td>\n",
       "      <td>0.9591829686</td>\n",
       "      <td>9.3623955760</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.99</th>\n",
       "      <td>0.0007007620</td>\n",
       "      <td>0.0025400495</td>\n",
       "      <td>0.0049225948</td>\n",
       "      <td>0.9918365937</td>\n",
       "      <td>10.5014208421</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       $\\lambda_0$   $\\lambda_1$   $\\lambda_2$   $\\lambda_3$       Entropie\n",
       "0.00  0.0700761971  0.2540049547  0.4922594755  0.1836593727   7.4685164076\n",
       "0.10  0.0630685773  0.2286044593  0.4430335280  0.2652934354   7.4700973606\n",
       "0.20  0.0560609576  0.2032039638  0.3938075804  0.3469274982   7.4969927508\n",
       "0.30  0.0490533379  0.1778034683  0.3445816329  0.4285615609   7.5467698528\n",
       "0.40  0.0420457182  0.1524029728  0.2953556853  0.5101956236   7.6204790432\n",
       "0.50  0.0350380985  0.1270024774  0.2461297378  0.5918296864   7.7223233454\n",
       "0.60  0.0280304788  0.1016019819  0.1969037902  0.6734637491   7.8610268679\n",
       "0.70  0.0210228591  0.0762014864  0.1476778427  0.7550978118   8.0541442064\n",
       "0.80  0.0140152394  0.0508009909  0.0984518951  0.8367318745   8.3418290591\n",
       "0.90  0.0070076197  0.0254004955  0.0492259476  0.9183659373   8.8510590988\n",
       "0.95  0.0035038099  0.0127002477  0.0246129738  0.9591829686   9.3623955760\n",
       "0.99  0.0007007620  0.0025400495  0.0049225948  0.9918365937  10.5014208421"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h2>TEXTEN1.txt - DECREASE</h2>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$\\lambda_0$</th>\n",
       "      <th>$\\lambda_1$</th>\n",
       "      <th>$\\lambda_2$</th>\n",
       "      <th>$\\lambda_3$</th>\n",
       "      <th>Entropie</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1.0</th>\n",
       "      <td>0.0700761971</td>\n",
       "      <td>0.2540049547</td>\n",
       "      <td>0.4922594755</td>\n",
       "      <td>0.1836593727</td>\n",
       "      <td>7.4685164076</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.9</th>\n",
       "      <td>0.0716527632</td>\n",
       "      <td>0.2597195289</td>\n",
       "      <td>0.5033342724</td>\n",
       "      <td>0.1652934354</td>\n",
       "      <td>7.4723589030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.8</th>\n",
       "      <td>0.0732293294</td>\n",
       "      <td>0.2654341031</td>\n",
       "      <td>0.5144090693</td>\n",
       "      <td>0.1469274982</td>\n",
       "      <td>7.4780701337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.7</th>\n",
       "      <td>0.0748058956</td>\n",
       "      <td>0.2711486773</td>\n",
       "      <td>0.5254838662</td>\n",
       "      <td>0.1285615609</td>\n",
       "      <td>7.4858966602</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.6</th>\n",
       "      <td>0.0763824618</td>\n",
       "      <td>0.2768632514</td>\n",
       "      <td>0.5365586631</td>\n",
       "      <td>0.1101956236</td>\n",
       "      <td>7.4961755430</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.5</th>\n",
       "      <td>0.0779590280</td>\n",
       "      <td>0.2825778256</td>\n",
       "      <td>0.5476334600</td>\n",
       "      <td>0.0918296864</td>\n",
       "      <td>7.5093852221</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.4</th>\n",
       "      <td>0.0795355942</td>\n",
       "      <td>0.2882923998</td>\n",
       "      <td>0.5587082569</td>\n",
       "      <td>0.0734637491</td>\n",
       "      <td>7.5262426076</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.3</th>\n",
       "      <td>0.0811121604</td>\n",
       "      <td>0.2940069739</td>\n",
       "      <td>0.5697830538</td>\n",
       "      <td>0.0550978118</td>\n",
       "      <td>7.5479128503</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.2</th>\n",
       "      <td>0.0826887266</td>\n",
       "      <td>0.2997215481</td>\n",
       "      <td>0.5808578507</td>\n",
       "      <td>0.0367318745</td>\n",
       "      <td>7.5765575585</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.1</th>\n",
       "      <td>0.0842652928</td>\n",
       "      <td>0.3054361223</td>\n",
       "      <td>0.5919326476</td>\n",
       "      <td>0.0183659373</td>\n",
       "      <td>7.6173517820</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.0</th>\n",
       "      <td>0.0858418590</td>\n",
       "      <td>0.3111506965</td>\n",
       "      <td>0.6030074445</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>7.7042619372</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      $\\lambda_0$   $\\lambda_1$   $\\lambda_2$   $\\lambda_3$      Entropie\n",
       "1.0  0.0700761971  0.2540049547  0.4922594755  0.1836593727  7.4685164076\n",
       "0.9  0.0716527632  0.2597195289  0.5033342724  0.1652934354  7.4723589030\n",
       "0.8  0.0732293294  0.2654341031  0.5144090693  0.1469274982  7.4780701337\n",
       "0.7  0.0748058956  0.2711486773  0.5254838662  0.1285615609  7.4858966602\n",
       "0.6  0.0763824618  0.2768632514  0.5365586631  0.1101956236  7.4961755430\n",
       "0.5  0.0779590280  0.2825778256  0.5476334600  0.0918296864  7.5093852221\n",
       "0.4  0.0795355942  0.2882923998  0.5587082569  0.0734637491  7.5262426076\n",
       "0.3  0.0811121604  0.2940069739  0.5697830538  0.0550978118  7.5479128503\n",
       "0.2  0.0826887266  0.2997215481  0.5808578507  0.0367318745  7.5765575585\n",
       "0.1  0.0842652928  0.3054361223  0.5919326476  0.0183659373  7.6173517820\n",
       "0.0  0.0858418590  0.3111506965  0.6030074445  0.0000000000  7.7042619372"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h2>TEXTCZ1.txt - INCREASE</h2>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$\\lambda_0$</th>\n",
       "      <th>$\\lambda_1$</th>\n",
       "      <th>$\\lambda_2$</th>\n",
       "      <th>$\\lambda_3$</th>\n",
       "      <th>Entropie</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.00</th>\n",
       "      <td>0.1402643298</td>\n",
       "      <td>0.4289372102</td>\n",
       "      <td>0.2446182953</td>\n",
       "      <td>0.1861801647</td>\n",
       "      <td>10.2208745744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.10</th>\n",
       "      <td>0.1262378968</td>\n",
       "      <td>0.3860434892</td>\n",
       "      <td>0.2201564658</td>\n",
       "      <td>0.2675621482</td>\n",
       "      <td>10.2253541802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.20</th>\n",
       "      <td>0.1122114638</td>\n",
       "      <td>0.3431497682</td>\n",
       "      <td>0.1956946363</td>\n",
       "      <td>0.3489441317</td>\n",
       "      <td>10.2548136581</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.30</th>\n",
       "      <td>0.0981850309</td>\n",
       "      <td>0.3002560472</td>\n",
       "      <td>0.1712328067</td>\n",
       "      <td>0.4303261153</td>\n",
       "      <td>10.3069213223</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.40</th>\n",
       "      <td>0.0841585979</td>\n",
       "      <td>0.2573623261</td>\n",
       "      <td>0.1467709772</td>\n",
       "      <td>0.5117080988</td>\n",
       "      <td>10.3827127929</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.50</th>\n",
       "      <td>0.0701321649</td>\n",
       "      <td>0.2144686051</td>\n",
       "      <td>0.1223091477</td>\n",
       "      <td>0.5930900823</td>\n",
       "      <td>10.4862783639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.60</th>\n",
       "      <td>0.0561057319</td>\n",
       "      <td>0.1715748841</td>\n",
       "      <td>0.0978473181</td>\n",
       "      <td>0.6744720659</td>\n",
       "      <td>10.6260798434</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.70</th>\n",
       "      <td>0.0420792989</td>\n",
       "      <td>0.1286811631</td>\n",
       "      <td>0.0733854886</td>\n",
       "      <td>0.7558540494</td>\n",
       "      <td>10.8190774891</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.80</th>\n",
       "      <td>0.0280528660</td>\n",
       "      <td>0.0857874420</td>\n",
       "      <td>0.0489236591</td>\n",
       "      <td>0.8372360329</td>\n",
       "      <td>11.1038523521</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.90</th>\n",
       "      <td>0.0140264330</td>\n",
       "      <td>0.0428937210</td>\n",
       "      <td>0.0244618295</td>\n",
       "      <td>0.9186180165</td>\n",
       "      <td>11.6013904191</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.95</th>\n",
       "      <td>0.0070132165</td>\n",
       "      <td>0.0214468605</td>\n",
       "      <td>0.0122309148</td>\n",
       "      <td>0.9593090082</td>\n",
       "      <td>12.0936926732</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.99</th>\n",
       "      <td>0.0014026433</td>\n",
       "      <td>0.0042893721</td>\n",
       "      <td>0.0024461830</td>\n",
       "      <td>0.9918618016</td>\n",
       "      <td>13.1659710379</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       $\\lambda_0$   $\\lambda_1$   $\\lambda_2$   $\\lambda_3$       Entropie\n",
       "0.00  0.1402643298  0.4289372102  0.2446182953  0.1861801647  10.2208745744\n",
       "0.10  0.1262378968  0.3860434892  0.2201564658  0.2675621482  10.2253541802\n",
       "0.20  0.1122114638  0.3431497682  0.1956946363  0.3489441317  10.2548136581\n",
       "0.30  0.0981850309  0.3002560472  0.1712328067  0.4303261153  10.3069213223\n",
       "0.40  0.0841585979  0.2573623261  0.1467709772  0.5117080988  10.3827127929\n",
       "0.50  0.0701321649  0.2144686051  0.1223091477  0.5930900823  10.4862783639\n",
       "0.60  0.0561057319  0.1715748841  0.0978473181  0.6744720659  10.6260798434\n",
       "0.70  0.0420792989  0.1286811631  0.0733854886  0.7558540494  10.8190774891\n",
       "0.80  0.0280528660  0.0857874420  0.0489236591  0.8372360329  11.1038523521\n",
       "0.90  0.0140264330  0.0428937210  0.0244618295  0.9186180165  11.6013904191\n",
       "0.95  0.0070132165  0.0214468605  0.0122309148  0.9593090082  12.0936926732\n",
       "0.99  0.0014026433  0.0042893721  0.0024461830  0.9918618016  13.1659710379"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h2>TEXTCZ1.txt - DECREASE</h2>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$\\lambda_0$</th>\n",
       "      <th>$\\lambda_1$</th>\n",
       "      <th>$\\lambda_2$</th>\n",
       "      <th>$\\lambda_3$</th>\n",
       "      <th>Entropie</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1.0</th>\n",
       "      <td>0.1402643298</td>\n",
       "      <td>0.4289372102</td>\n",
       "      <td>0.2446182953</td>\n",
       "      <td>0.1861801647</td>\n",
       "      <td>10.2208745744</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.9</th>\n",
       "      <td>0.1434732017</td>\n",
       "      <td>0.4387501439</td>\n",
       "      <td>0.2502145062</td>\n",
       "      <td>0.1675621482</td>\n",
       "      <td>10.2240601916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.8</th>\n",
       "      <td>0.1466820736</td>\n",
       "      <td>0.4485630775</td>\n",
       "      <td>0.2558107172</td>\n",
       "      <td>0.1489441317</td>\n",
       "      <td>10.2291426565</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.7</th>\n",
       "      <td>0.1498909455</td>\n",
       "      <td>0.4583760112</td>\n",
       "      <td>0.2614069281</td>\n",
       "      <td>0.1303261153</td>\n",
       "      <td>10.2363751923</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.6</th>\n",
       "      <td>0.1530998174</td>\n",
       "      <td>0.4681889448</td>\n",
       "      <td>0.2670031390</td>\n",
       "      <td>0.1117080988</td>\n",
       "      <td>10.2461075907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.5</th>\n",
       "      <td>0.1563086893</td>\n",
       "      <td>0.4780018785</td>\n",
       "      <td>0.2725993499</td>\n",
       "      <td>0.0930900823</td>\n",
       "      <td>10.2588433397</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.4</th>\n",
       "      <td>0.1595175612</td>\n",
       "      <td>0.4878148121</td>\n",
       "      <td>0.2781955608</td>\n",
       "      <td>0.0744720659</td>\n",
       "      <td>10.2753514558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.3</th>\n",
       "      <td>0.1627264331</td>\n",
       "      <td>0.4976277457</td>\n",
       "      <td>0.2837917718</td>\n",
       "      <td>0.0558540494</td>\n",
       "      <td>10.2969158646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.2</th>\n",
       "      <td>0.1659353050</td>\n",
       "      <td>0.5074406794</td>\n",
       "      <td>0.2893879827</td>\n",
       "      <td>0.0372360329</td>\n",
       "      <td>10.3260146914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.1</th>\n",
       "      <td>0.1691441769</td>\n",
       "      <td>0.5172536130</td>\n",
       "      <td>0.2949841936</td>\n",
       "      <td>0.0186180165</td>\n",
       "      <td>10.3689846354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.0</th>\n",
       "      <td>0.1723530488</td>\n",
       "      <td>0.5270665467</td>\n",
       "      <td>0.3005804045</td>\n",
       "      <td>0.0000000000</td>\n",
       "      <td>10.4895561069</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      $\\lambda_0$   $\\lambda_1$   $\\lambda_2$   $\\lambda_3$       Entropie\n",
       "1.0  0.1402643298  0.4289372102  0.2446182953  0.1861801647  10.2208745744\n",
       "0.9  0.1434732017  0.4387501439  0.2502145062  0.1675621482  10.2240601916\n",
       "0.8  0.1466820736  0.4485630775  0.2558107172  0.1489441317  10.2291426565\n",
       "0.7  0.1498909455  0.4583760112  0.2614069281  0.1303261153  10.2363751923\n",
       "0.6  0.1530998174  0.4681889448  0.2670031390  0.1117080988  10.2461075907\n",
       "0.5  0.1563086893  0.4780018785  0.2725993499  0.0930900823  10.2588433397\n",
       "0.4  0.1595175612  0.4878148121  0.2781955608  0.0744720659  10.2753514558\n",
       "0.3  0.1627264331  0.4976277457  0.2837917718  0.0558540494  10.2969158646\n",
       "0.2  0.1659353050  0.5074406794  0.2893879827  0.0372360329  10.3260146914\n",
       "0.1  0.1691441769  0.5172536130  0.2949841936  0.0186180165  10.3689846354\n",
       "0.0  0.1723530488  0.5270665467  0.3005804045  0.0000000000  10.4895561069"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAzUAAAGMCAYAAADjtox4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2373581bf28>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<h1>Pokrytí</h1>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TEXTEN1.txt</th>\n",
       "      <th>TEXTCZ1.txt</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Uni-gram</th>\n",
       "      <td>0.955602</td>\n",
       "      <td>0.864557</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Bi-gram</th>\n",
       "      <td>0.690615</td>\n",
       "      <td>0.420479</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Tri-gram</th>\n",
       "      <td>0.340300</td>\n",
       "      <td>0.175450</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          TEXTEN1.txt  TEXTCZ1.txt\n",
       "Uni-gram     0.955602     0.864557\n",
       "Bi-gram      0.690615     0.420479\n",
       "Tri-gram     0.340300     0.175450"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAk0AAAGGCAYAAABmPbWyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAPYQAAD2EBqD+naQAAIABJREFUeJzt3XucVVX9//HXh4vhOCoKClKSUhr69ZJOmmV2Ec00S9PU\nwBTNQIVUINNvFikQaqXmXaIsLyjIr/Lahb6gpl+1G3jJS5aKaKUokwjDMAqyfn/sPXyHwwxs5sI5\nA6/n43EecNZZa+/PzGMz82btvdeOlBKSJElasy7lLkCSJKkzMDRJkiQVYGiSJEkqwNAkSZJUgKFJ\nkiSpAEOTJElSAYYmSZKkAgxNkiRJBRiaJKkTiohNIuJbEXFwuWuRNhaGJkmdTkR8JiIejYilEfFO\nRGyxHvf9YkTctb72twYXAKcDj5a5DmmjYWiS1CYRsUNEXB0Rz0bEkvz1VN62ewfsb2vgNqAeGAGc\nACxp7/2swXp59lRE7BIR50dE/2Y++yAwCjgupbRgfdQjCbqVuwBJnVdEHA5MA5YBtwCPAyuAgcBR\nwGkRsWNK6eV23O0+QDXw7ZTSfe243UqzK3A+cB/wUmNjRHQBfgJckFJ6qEy1SRslQ5OkVomIAcBU\nYC4wKKX0Wsnn55LNBK1Yy3aqUkr167DrPvmfb67DmLKLiK5Al5TSsqJDaGZWK6W0AvhQe9YmqRhP\nz0lqrXOBKuDk0sAE2S/3lNLVKaV/NbZFxA0RsTgiBkTEryNiETAl/+xjETE9IuZFRENEvBQRl0VE\njybj7wNuyN/+JSJWRMRPWyowIi7I+3wg3/abEbEgIi6PiHeV9O0aEWMj4rl8/3MjYmJEbLK2b0RE\nDI2IZRHxvfz9e/P9jomIsyLiOaAB2Dci6iLih81s490RsTwizo2IocD0/KP78229ExEfz/veHxH3\nrq0uSe3LmSZJrfVZ4LmU0l/WYUwi+7kzA3gQ+DrZtUkAxwCbAtcCtcC+wBnAu4Hj8j7fBZ4FhgHf\nBl4Enl/L/iALIHOB/wb2A84EegInNel7PXBi3vcS4MPAN8lONR7d0g4iYjhwHfDdlNL5JR9/BXgX\n8CPgLWAecDtwXESMSSk1nUkakv85BdgEuDL/+r8L/C3/7JmSr0vSemRokrTOImJzoB9ZACj9bEtW\n/dmyJKXU0OT9JsBtKaVvlww9J6X0VpP3P4mI54GJEfGelNI/U0qzIuI9ZKHptymlOQVLfj6ldFT+\n9+siYjFwekRcklJ6MiL2IAtMk1NKp+X9JkXE68DXI+ITKaXfN/O1ngn8kOz6qoua2e+7gfellP7T\nZMxNZAHpYOB3TfoeDzzQODMXEQ+ShaaZKaUHCn6dkjqQp+cktUbjLf51zXx2P/B6k9eIZvpMKm1o\nGpgioioiegGPkP2c2qsNtSbgmpK2q8iuGTosf//ZvF/pabNL836fLd1oRHwDuBz4RguBCeDnTQNT\nbibwCllIatzWbsAewM1r+2IklY8zTZJaY3H+Z3Uznw0HNie7YHtKM58vTyn9s7QxIrYHJgCfA7Zq\n8lECtmxTtfBcyfvnyS5Q3yF/3z9/v0q/lNL8iFgIvLdk/CeBw4GLU0qXrWG/L5Y2pJRSRNxCdmdh\nj3wW7nhgKfDzIl+MpPJwpknSOkspLSKbLdmtmc/+nFK6F3iYbJam1FulDflt9DOBQ4GLgCOAg4Ch\n+Tba+2dVS9cEFb1W6Emy64xOiIgd1tBvaQvtN5EFyyPz94OBu1NKi1voL6kCGJoktdavgPdHRHvc\n/r47sBMwJqV0SUrp7jx4vdIO2ybfdlPvJ/v5Nzd/Py9/v0q/iNiW7ILxeSXjF5CFuuXArIjouy7F\npJSeIlvJ+/iIOIBspqv01JwXe0sVxtAkqbW+TzaT8tM8XJRal58v77QwZhRtDw8BjCxpOzPf7m/z\n97/O+40q6ff1vN+vSjeaUvo3WXDaFPifiNiqtM9a3Awcku9zQZNaGi3Ja+q5jtuV1EG8pklSq6SU\nnouIIcCtwLP5dTqPk/2i35HsDrF3gNWuX2rG38iuM7o0vztuEdlt/i0FhuZO+63JjhFxJ1kw+SjZ\nNURTUkp/zb+WJyLiRmB4Hn5+T7bkwInAL5u7cy4f93xEfJrs4vffRcSB63CK7Vay4HkkcG1K6Z2S\nzx8j+/6dGxE9yU5rzvKxKVL5ONMkqdVSSneRnVq7hewW+suBy4DPA3cDe6eUppcOa2Y7y8kurH6U\nbC2l75Ctx3RiS7telzLJ1nl6i+x6qUPJ1kD6akm/U8geW/IhsrvoPglMJLveqHR7K/efUnoy3+ZO\nwF1NFs1cpd9qRWULgjYuObDaBfMppfnAqcC2ZI9NuZXs0SpN65C0HsWqa6tJ0oYjIs4nC2DbNHPr\nf9lFxC+B3VJKO5e7FklrVxEzTRFxQETcFRH/yh8X8PkCYz4ZEbPzxx38PX/sgCR1ChGxHdn6TzeV\nuxZJxVREaAI2Izt/P4ICU875Lb73ALOAPYEryFYPPrjjSpSktouIHSLiy2QPO34bmFzmkiQVVBEX\ngqeUfkt+50hEFLnA83TghZTSOfn7ZyPiY8Bo4H86pkpJahefAH5GtvDlic097FhSZaqUmaZ1tR/Z\nQnhNzQA+UoZaJFWolNK4lFLXSrqeKaV0Y0qpS0ppQEpptWf3SapcnTU09QXml7TNB7ZocueKJElS\nu6mI03PrQ/7wz0PIpsQb1txbkiRtRHqQPYtyRkqptqVOnTU0vUr2MNCm+gCLmj4pvcQhZGvJSJIk\nNed4sjXRmtVZQ9MjZIvJNfXpvL0lLwJMmTKFXXbZpYPK6pxGjx7ND3/4w3KXIRXmMavOyOO2cj3z\nzDN8+ctfhjwrtKQiQlNEbEb2AM3GO+cGRMSewH9SSi9HxEVAv5RS41pMk4CREfE94KfAIOCLwGFr\n2E0DwC677MLee+/dEV9Gp7Xlllv6PVGn4jGrzsjjtlNY4+U7lXIh+IfIHp8wm2ydpkuBOcC4/PO+\nwPaNnVNKL5ItCncQ2fpOo4FTUkqld9RJkiS1i4qYacofhtligEspndxM2wNATUfWJUmS1KhSZpok\nSZIqmqFJDB5c+hB3qbJ5zKoz8rjt/Cri9JzKy3/I6mw8ZjuHl156iQULFpS7jIrxgQ98gDlz5pS7\njI1S79696d+/f5u3Y2iSJLW7l156iV122YX6+vpylyJRVVXFM8880+bgZGiSJLW7BQsWUF9f79p4\nKrvGNZgWLFhgaJIkVS7XxtOGxAvBJUmSCtjoZpreeOMNL0ws0aNHD6qrq8tdhiRJFW2jC0333Xor\nz997b7nLqCjdevXi2OHDDU6SJK3BRhea9nnXu9i/V69yl1ExFtbXc29tLQ0NDYYmSZLWYKMLTdU9\netDbcLCqpUvLXYEkSRXPC8ElSetVXV0dCxYsKPurrq5unWvv0qXLWl9du3blgQceYN68eWvs9/3v\nfx+A119/nV69enHQQQettr/ly5ez++67M2DAAOrr6ztk/wCf/OQn6dKlC0ccccRqNTRu57LLLlul\nfeLEiRxxxBH07duXLl26MH78+MLfx0ceeYRx48axaNGiwmOaWrp0KePGjeOBBx5o1fjW2uhmmiRJ\n5VNXV8f0yZNZXltb7lJadT3nlClTVnl/4403MnPmTKZMmUJKaWV704U9hwwZwmGHHbbatvbaay8A\nttlmG773ve9x6qmncvPNN3PCCSes7HPJJZfw9NNPc88991BVVdUh+weICCKCe+65h0cffXSVz1oy\nduxYtttuO/bee29mzJix1v5NPfzww4wfP56TTz6ZLbbYYp3GAtTX1zNu3Dgigo9//OPrPL61DE2S\npPWmoaGB5bW1HLjppvSsqipbHa29nnPIkCGrvH/kkUeYOXNms4/2mTdvHgB77733auNKffWrX+Wm\nm27i7LPP5vDDD2errbZi7ty5TJgwgaOPPppDDz20Q/cP0L9/fxYvXsy4ceO444471tr/xRdfpH//\n/tTW1rLNNtustX9TTQNea7R1fGt5ek6StN71rKqid3V12V7lDGwtmTRpEgsXLuTss88GYMSIEXTv\n3p0rrrhivex/8803Z/To0dx111089thja+3f2tW1x40bxznnnAPADjvssPKU4ksvvcQNN9xAly5d\nuOGGG1YZc+GFF9KlSxd++9vfMm/ePLbddlsiggsuuGDl6cZ1OT3YWs40SZLUgerr66lt5nRkz549\n6dq168r3u+66K2effTYXX3wx1dXVzJgxg6uuuorttttuvewf4KyzzuKyyy7jggsuKDTb1BpHH300\nf//735k2bRpXXHEFvfI72rfZZhtOOukkfvnLXzJmzBgOPvhg3v3ud/PXv/6V8ePHM2zYMD7zmc9Q\nX1/PpEmTOO200zjqqKM46qijANhjjz06pN6mnGmSJKkDnX/++WyzzTarvLbddltmz569Wt+xY8ey\n4447ctVVV7HPPvswYsSI9br/6upqRo0axd13311otqk1dtttt5WP1jniiCMYMmQIQ4YMYdNNNwXg\nxz/+MV26dOGUU05h2bJlDB06lH79+nHppZcC2cN3jz76aCALSo3jd9tttw6ptylnmiRJ6kDDhw/n\nmGOOWa191113Xa1tk002YcsttyQiOPDAA4mI9bp/yGabLr/8csaNG8ftt9/e5v2vqz59+nDNNdcw\nePBgDjjgAJ544glmzpxZEWsJGpokSepAO+20EwceeGChvpdffjmPPvoou+++O1deeSXDhg1jwIAB\n623/AFtssQWjRo3iggsu4PHHH6dnz55t2n9rHHfccUyZMoVf/epXnHrqqXzyk59c7zU0x9NzkiRV\ngJdffpkLLriAo446it/97nd0796dkSNHlqWWs846iy233JJx48aVZf//+c9/+Mtf/kJE8PTTT5el\nhuYYmiRJqgBf+9rXiAiuvPJK+vTpw8SJE5kxYwbTp09f77U0zjbdeeedPProo+2+/bWddhwxYgR1\ndXVcdNFFPPjgg1x++eXrNL6jGJokSSqz22+/nbvvvpsJEybQr18/IAsONTU1jBkzplWrl7fVqFGj\n2HLLLRk/fnybQkptbS3PPvssS5s8smuzzTYDYOHChav1//nPf8706dP53ve+xznnnMOXvvQlvv3t\nb/Pcc8+t7FOVLxnR3PiO5DVNkiR1oNmzZ3PLLbes1v6+972P/fbbj8WLF3PmmWdSU1PDGWecsfLz\niGDSpEl8+MMf5rzzzuPKK6/skP23ZIsttuCss85aufJ2qSlTpjBv3jyWLFkCwO9//3smTpwIwIkn\nnsj2228PwFVXXcX48eO5//77V67eXVNTQ0qJ8847jy996Ut0796dz3/+8yxevJjTTz+dQYMGrbxz\n8Oqrr+a+++5j6NChPPTQQwD06NGDXXfdldtuu42ddtqJrbfemt12243/+q//atX3qChDkyRpvVuY\nP+JjQ9j/mmZhIoJp06Yxbdq01T4bOnQo++23H9/5znd49dVXufPOO1fbVk1NDSNGjOC6667j5JNP\nbvbxJm3d/5q2M2rUKK644grefPPN1T67/vrrVz77LSK4//77uf/++wE44IADVoamxke0NPWhD32I\n7373u0yaNIkZM2awYsUK5s6dy5gxY1i+fDk/+9nPVvbdeuutmTx5MkceeSSXXHLJysU/r7/+es44\n4wzGjBnD22+/zfnnn9/hoSnKtRT5+hYRewOzZ44ezaCBA8tdTsVYUFfHL2trOWr0aHr37l3uciRt\nIObMmUNNTQ2zZ89euSYPdP5nz6nzaelYbK4PUJNSmtPStpxpkiStN9XV1Rw7fDgNDQ3lLoUePXoY\nmLRODE2SpPWqurrasKJOybvnJEmSCjA0SZIkFWBokiRJKsDQJEmSVIChSZIkqQBDkyRJUgGGJkmS\npAIMTZIkSQW4uKVoePttaivgkQaVxJWCJUmlDE0bubq33uLBR+cyd9KvqararNzlVIxevboxfPix\nBidJ0kqGpo3cW8uWsWRpdzbd9FP06vWecpdTEerrF1Jbey8NDQ2GJqkD1NXV+ew5dUqGJgHQo0dP\nqqt7l7uMirF0abkrkDZMdXV1TJ48ndra5eUupVUzyl26rP1S4Ijgvvvu473vfS877rhji/0uvvhi\nzjnnHF5//XUGDhzIXnvtxcyZM1fps3z5cvbaay+WLFnCk08+WajWxv1//OMfB2DRokVceuml3HHH\nHbzwwgu88847vP/97+fwww/nzDPPpG/fvsyaNYuDDz54rdv+3//9Xz760Y8CMHHiRP70pz/xxz/+\nkddee43vfve7nHfeeWvdBsBDDz3EzJkz+frXv96q4LpkyRIuueQSBg0axMc+9rF1Ht9ahiZJ0nrT\n0NBAbe1yNt30QKqqepatjtbOKE+ZMmWV9zfeeCMzZ85kypQppJRWtu+yyy7U19cDMGTIEA477LDV\ntrXXXnsBsM022/C9732PU089lZtvvpkTTjhhZZ9LLrmEp59+mnvuuYeqqqp12j/Ac889x0EHHcQr\nr7zCsccey+mnn063bt144okn+PGPf8xdd93Fk08+yW677bbathstWrSIUaNGscUWW7DTTjsB8M47\n7zB27Fj69evH3nvvzYwZMwp/DyELX+PHj2fYsGGtCk11dXWMGzeO7t27G5okSRu2qqryz263ZkZ5\nyJAhq7x/5JFHmDlzJoMHD16t77x58wDYe++9VxtX6qtf/So33XQTZ599NocffjhbbbUVc+fOZcKE\nCRx99NEceuih67z/5cuX84UvfIE33niDBx98kH333XeVzydOnMgPfvADAPr06dNijV/60pdYtmwZ\nP/vZz9hmm20A6Nq1K/PmzWP77bdn/vz5bLfddmv8+ko1DXit0dbxreWSA5IkVYBJkyaxcOFCzj77\nbABGjBhB9+7dueKKK1q1vdtuu42nnnqK73znO6sFJoDNN9+c8ePHr3EbkydPZvr06Zx55pl89rOf\nXeWz7bffvlV1jR07duVpvPe85z106dKFrl278u9//5uf/OQndOnSZbVZr/Hjx9OlSxdmzpzJ888/\nT79+/YgIvv3tb9OlSxe6dOnChRde2Kp61oUzTZIkdaD6+vpml3Xp2bMnXbt2Xfl+11135eyzz+bi\niy+murqaGTNmcNVVV63zLE6ju+66i4jgy1/+cqvGP/nkk4wePZqamhq+//3vt2obzTn22GN57rnn\nmD59OldffTU9e2anabfeemu++tWv8stf/pKzzjqLQYMGsd122/HYY48xceJETj/9dA466CCWLFnC\nNddcw8iRIznmmGM44ogjAPjgBz/YbjW2xJkmSZI60Pnnn88222yzymvbbbdl9uzZq/UdO3YsO+64\nI1dddRX77LMPI0aMaPV+//a3v9GrVy/69OmzzmOXLl3KscceS7du3Zg2bRrdu3dvdR2ldt9995UB\n58gjj2TIkCEMGTKEHj16APCTn/yElBLDhg1j2bJlDB06lP79+688lbjZZptx1FFHAbDnnnuuHL/r\nrru2W40tcaZJkqQONHz4cI455pjV2pv7Jb/JJpuw5ZZbEhEceOCBRESr97to0SI233zzVo0dOXIk\nzz77LDfddBPve9/7Wl1Da/Tr14+rrrqKE044gQMOOICnnnqK++67j6qqqvVaR3MMTZIkdaCddtqJ\nAw88sFDfyy+/nEcffZTdd9+dK6+8kmHDhjFgwIBW7XeLLbbglVdeWedxt9xyCzfccANDhw7l+OOP\nb9W+2+r4449nypQpzJgxg5EjR3LAAQeUpY5Snp6TJKkCvPzyy1xwwQUcddRR/O53v6N79+6MHDmy\n1dsbOHAgtbW1zJ8/v/CYf/zjH5x++ukMHDiQa665ptX7bqsFCxYwZ84cIoKnnnqqbHWUMjRJklQB\nvva1rxERXHnllfTp04eJEycyY8YMpk+f3qrtfe5znyOl1OL6S6Xefvttjj32WJYvX860adM69HTY\n2k47nn766SxdupQLL7yQ+++/n6uvvnqdxncUQ5MkSWV2++23c/fddzNhwgT69esHZEsO1NTUMGbM\nGOrq6tZ5m8cddxy77rorEyZM4M9//vNqny9atIixY8eufD9mzBieeOIJfvCDH7DHHnu0/ospUVtb\ny7PPPrvKo3M22yx71unChQtX6z9t2jR+8YtfcMkll3DuuefyxS9+kW9+85u88MILhcZ3JK9pkiSt\nd/X16/eXXTn3P3v2bG655ZbV2t/3vvex3377sXjxYs4880xqamo444wzVn4eEUyaNIkPf/jDnHfe\neVx55ZXrtN/u3btz++238+lPf5r999+f4447jv33359u3brx5JNPcuutt9KnTx8mTJjAPffcw7XX\nXsv2229Pz549m60XYP/992eHHXYA4Oabb+all15i8eLFANx///0rF5086aSTePe73w3AD3/4Qy68\n8MJVHsFSU1NDSolvfvObHHPMMXTv3p0jjzySN954g6997WsccsghDB8+HIBrr72W//qv/+Kkk07i\ngQceALLQtPPOOzN16lQGDBjAVlttxR577LFyJfSOYmiSJK03PXr0oFevbtTW3lv2Zzz26tVt5W3u\nbbGmU0URwbRp05g2bdpqnw0dOpT99tuP73znO7z66qvceeedq22rpqaGESNGcN1113HyySevfPRK\n0f3vtNNOPPbYY1x22WXccccd3HHHHaxYsYIBAwYwbNgwRo0aBcAf//hHIoJ//vOfnHjiiS1u7+ab\nb14ZmiZPnszDDz+8soZZs2Yxa9YsAD71qU+tDE0RsVqN++23H+PGjWPy5Mn8+te/ZsWKFbz88suM\nGDGCFStWcP3116/s27t3byZNmsQXv/hFLr/88pU1//SnP+Wss85i9OjRvP3220yYMKHDQ1OUayny\nUhExEjgb6As8DpyRUlp9PvH/+h8PfAPYCXgT+A3wjZTSf1rovzcwe+bo0QwaOLC9y++0nn31VUbP\neIb9DrmIvn13KHc5FaGubgG1tb9k9Oij6N3bhxhLrTFnzhxqamqYPXs2e++99yqf1dXVrXKqplx6\n9OjRqueeqXNZ07FY2geoSSnNaWlbFTHTFBHHAZcCw4E/AaOBGRGxc0ppQTP99wduBM4C7gHeDfwI\nmAx8cX3VLUlad9XV1YYVdUqVciH4aOBHKaWbUkp/A04D6oGvtNB/P2BuSumalNK8lNLDZKFp9Yfr\nSJIktYOyh6aI6A7UALMa21J2znAm8JEWhj0CbB8Rh+bb6AMcA/yqY6uVJEkbq7KHJqA30BUoXX1r\nPtn1TavJZ5a+DNwWEW8DrwBvAF/rwDolSdJGrCKuaVpXEbErcAVwAfA7YDvgErJTdF9d09ixd97J\nFbNmrdI2eJ99GLyvZ/YkSdrQTZ06lalTp67S9uabbxYaWwmhaQHwDlD6GOY+wKstjPlv4KGU0mX5\n+ycjYgTwYER8K6XU4prxE444wrvnJEnaSA0ePJjBgwev0tbk7rk1KvvpuZTSMmA2MKixLbIFHQYB\nD7cwrApYXtK2AkhAedZWlyRJG7Syh6bcZcCwiDgxIgYCk8iC0Q0AEXFRRNzYpP/dwNERcVpE7Jgv\nQXAF8MeUUkuzU5IkSa1WCafnSClNj4jewHiy03KPAYeklF7Pu/QFtm/S/8aIqAZGkl3LtJDs7rv/\nXq+FS5LW6Jlnnil3CdrItecxWBGhCSCldC1wbQufndxM2zXANR1dlyRp3fXu3Zuqqiq+/OUvl7sU\niaqqqnZ5wkPFhCZJ0oajf//+PPPMMyxYsNpDHaT1rnfv3vTv37/N2zE0SZI6RP/+/dvlF5VUKSrl\nQnBJkqSKZmiSJEkqwNAkSZJUgKFJkiSpAEOTJElSAYYmSZKkAgxNkiRJBRiaJEmSCjA0SZIkFWBo\nkiRJKsDQJEmSVIChSZIkqQBDkyRJUgGGJkmSpAIMTZIkSQUYmiRJkgowNEmSJBVgaJIkSSrA0CRJ\nklSAoUmSJKkAQ5MkSVIBhiZJkqQCDE2SJEkFGJokSZIKMDRJkiQV0K3cBUhas7q6OhoaGspdRkXp\n0aMH1dXV5S5D0kbG0CRVsLq6OqZPnszy2tpyl1JRuvXqxbHDhxucJK1XhiapgjU0NLC8tpYDN92U\nnlVV5S6nIiysr+fe2loaGhoMTZLWK0OT1An0rKqitwHh/yxdWu4KJG2EvBBckiSpAEOTJElSAYYm\nSZKkAgxNkiRJBRiaJEmSCjA0SZIkFWBokiRJKsDQJEmSVIChSZIkqQBDkyRJUgGGJkmSpAIMTZIk\nSQUYmiRJkgowNEmSJBVgaJIkSSrA0CRJklSAoUmSJKkAQ5MkSVIBhiZJkqQCDE2SJEkFVExoioiR\nETE3IpZGxB8iYp+19N8kIiZGxIsR0RARL0TESeupXEmStJHpVu4CACLiOOBSYDjwJ2A0MCMidk4p\nLWhh2P8DtgFOBp4HtqOCQqAkSdqwVERoIgtJP0op3QQQEacBnwW+Any/tHNEfAY4ABiQUlqYN7+0\nnmqVJEkbobLPzEREd6AGmNXYllJKwEzgIy0M+xzwF+DciPhnRDwbET+IiB4dXrAkSdooVcJMU2+g\nKzC/pH0+8IEWxgwgm2lqAI7Mt3EdsDVwSseUKUmSNmaVEJpaowuwAhiSUqoDiIgxwP+LiBEppbda\nGjj2zju5YtasVdoG77MPg/fdtyPrlSRJFWDq1KlMnTp1lbY333yz0NhKCE0LgHeAPiXtfYBXWxjz\nCvCvxsCUewYI4D1kF4Y3a8IRRzBo4MDWVytJkjqtwYMHM3jw4FXa5syZQ01NzVrHlv2appTSMmA2\nMKixLSIif/9wC8MeAvpFRFWTtg+QzT79s4NKlSRJG7Gyh6bcZcCwiDgxIgYCk4Aq4AaAiLgoIm5s\n0v9WoBb4WUTsEhEfJ7vL7vo1nZqTJElqrUo4PUdKaXpE9AbGk52Weww4JKX0et6lL7B9k/5LIuJg\n4Crgz2QB6jZg7HotXJIkbTQqIjQBpJSuBa5t4bOTm2n7O3BIR9clSZIElXN6TpIkqaIZmiRJkgow\nNEmSJBVgaJIkSSqgcGiKiE9ExKYdWYwkSVKlWpeZpg8A90VEL4CI2GJNr44pV5IkqTwKLzmQUpqc\nLdTN/cBx0f+VAAAXnUlEQVTuwEIgNdM18vau7VCfJElSRVindZry4PRo/vZAmg9NkiRJG5x1Xtwy\npfTn/K8vAC+nlFYJTvlz47ZfbaAkSVIn1pa75+YC2zTTvnX+mSRJ0gajLaGp8dqlUtVAQxu2K0mS\nVHHW+fRcRFyW/zUBEyKivsnHXYEPkz1wV5IkaYPRmgf27pX/GWR30b3d5LO3gceBS9pYlyRJUkVp\nzYXgnwKIiJ8BZ6WUFrV7VZIkSRWmLdc0PQAsa69CJEmSKllbQtPFwPyIuD4iPtpeBUmSJFWitoSm\ndwNDgd7A/RHxt4g4NyL6tk9pkiRJlaPVoSmltDyldHtK6QiyxSx/DBwPvBQRd0XEERHRllAmSZJU\nMdol1KSU5gP/CzwCrCC7q+5G4PmI+GR77EOSJKmc2hSaIqJPRJwdEU+RPch3C+DwlNKOZKfvppOF\nJ0mSpE6t1aEpIu4GXgZOIjs19+6U0uCU0kyAlNIS4FJ8Dp0kSdoAtGZxy0avAZ9IKT2yhj6vAzu2\nYR+SJEkVoVWhKSK6AzsAC9bUL6WUgHmt2YckSVIladXpuZTSMmCPdq5FkiSpYrXlQvApwCntVYgk\nSVIla8s1Td2Ar0TEQcBsYEnTD1NKY9pSmCRJUiVpS2jaDZiT/33ndqhFkiSpYrU6NKWUPtWehUiS\nJFWytqzT1GJoioiRrd2uJElSJWrLheC/jIia0saIOAu4qA3blSRJqjhtCU3fAH4TEQMbGyLi68B4\n4LNtLUySJKmStOWapp9ExNbAzIj4GHAccB5wWErpofYqUJIkqRK05e45Ukrfj4hewF+ArsAhKaU/\ntEtlkiRJFWSdQlNEnNlM87+AeuABYN+I2BcgpXRl28uTJEmqDOs60zS6hfZ3gP3zF0ACDE2SJGmD\nsU6hKaW0Y0cVIkmSVMnacvfcSpFrj21JkiRVojaFpog4MSL+CiwFlkbEExFxQvuUJkmSVDlaffdc\nRIwBJgBXA41LDHwMmBQRvVNKP2yH+iRJkipCW5YcOAM4PaV0U5O2uyLiKeACwNAkSZI2GG05Pbcd\n8HAz7Q/nn0mSJG0w2hKangOObab9OOAfbdiuJElSxWnL6bnzgdsi4uP83zVN+wODaD5MSZIkdVqt\nnmlKKf0C2BdYAByZvxYA+6aUbm+f8iRJkirDOs80RUQX4GzgCGAT4F5gWEppaTvXJkmSVDFaM9P0\nLeBCYDHZc+fOBK5pz6IkSZIqTWtC04nAiJTSZ1JKRwKfA47PZ6AkSZI2SK0JOv2B3zS+SSnNJHtA\nb7/2KkqSJKnStCY0dQMaStqWAd3bXo4kSVJlas2SAwHcEBFvNWnrQfb4lCWNDSmlo9panCQ1p+Ht\nt6mtrS13GRWnR48eVFdXl7sMaYPVmtB0YzNtU9paiCQVUffWWzz46FzmTvo1VVWblbucitKrVzeG\nDz/W4CR1kHUOTSmlkzuikIgYSbaUQV/gceCMlNKfC4zbH7gf+GtKae+OqE1S5Xhr2TKWLO3Oppt+\nil693lPucipGff1CamvvpaGhwdAkdZC2rAjebiLiOOBSYDjwJ2A0MCMidk4pLVjDuC3JZr5mAn3W\nR62SKkOPHj2pru5d7jIqylJXy5M6VKUsEzAa+FFK6aaU0t+A04B64CtrGTcJuAX4QwfXJ0mSNnJl\nD00R0R2oAWY1tqWUEtns0UfWMO5kYEdgXEfXKEmSVAmn53oDXYH5Je3zgQ80NyAidiJblfxjKaUV\nEdGxFUqSpI1eJYSmdZKvPH4LcH5K6fnG5qLjx955J1fMmrVK2+B99mHwvvu2X5GSJKkiTZ06lalT\np67S9uabbxYaWwmhaQHwDqtfyN0HeLWZ/psDHwI+GBGNz7zrAkREvA18OqV0f0s7m3DEEQwaOLDN\nRUuSpM5n8ODBDB48eJW2OXPmUFNTs9axZb+mKaW0DJgNDGpsi+x82yDg4WaGLAJ2Az4I7Jm/JgF/\ny//+xw4uWZIkbYQqYaYJ4DKyVcZn839LDlQBNwBExEVAv5TS0Pwi8aebDo6I14CGlNIz67VqSZK0\n0aiI0JRSmh4RvYHxZKflHgMOSSm9nnfpC2xfrvokSZIqIjQBpJSuBa5t4bM1rkKeUhqHSw9IkqQO\nVPZrmiRJkjoDQ5MkSVIBhiZJkqQCDE2SJEkFGJokSZIKMDRJkiQVYGiSJEkqwNAkSZJUgKFJkiSp\nAEOTJElSAYYmSZKkAirm2XOSpA1DXV0dDQ0N5S6j4vTo0YPq6upyl6E2MDRJktpNXV0d0ydPZnlt\nbblLqTjdevXi2OHDDU6dmKFJktRuGhoaWF5by4GbbkrPqqpyl1MxFtbXc29tLQ0NDYamTszQJElq\ndz2rquhtOFjV0qXlrkBt5IXgkiRJBRiaJEmSCjA0SZIkFWBokiRJKsDQJEmSVIChSZIkqQBDkyRJ\nUgGGJkmSpAIMTZIkSQUYmiRJkgowNEmSJBVgaJIkSSrA0CRJklSAoUmSJKkAQ5MkSVIBhiZJkqQC\nDE2SJEkFGJokSZIKMDRJkiQVYGiSJEkqwNAkSZJUgKFJkiSpAEOTJElSAYYmSZKkAgxNkiRJBRia\nJEmSCjA0SZIkFWBokiRJKsDQJEmSVIChSZIkqQBDkyRJUgGGJkmSpAIMTZIkSQUYmiRJkgowNEmS\nJBVgaJIkSSqgYkJTRIyMiLkRsTQi/hAR+6yh7xci4ncR8VpEvBkRD0fEp9dnvZIkaeNSEaEpIo4D\nLgXOB/YCHgdmRETvFoZ8HPgdcCiwN3AfcHdE7LkeypUkSRuhighNwGjgRymlm1JKfwNOA+qBrzTX\nOaU0OqV0SUppdkrp+ZTSt4B/AJ9bfyVLkqSNSdlDU0R0B2qAWY1tKaUEzAQ+UnAbAWwO/KcjapQk\nSSp7aAJ6A12B+SXt84G+BbfxDWAzYHo71iVJkrRSt3IX0FYRMQQYC3w+pbRgbf3H3nknV8yatUrb\n4H32YfC++3ZQhZIkqVJMnTqVqVOnrtL25ptvFhpbCaFpAfAO0KekvQ/w6poGRsSXgMnAF1NK9xXZ\n2YQjjmDQwIGtqVOSJHVygwcPZvDgwau0zZkzh5qamrWOLfvpuZTSMmA2MKixLb9GaRDwcEvjImIw\ncD3wpZTSbzu6TkmStHGrhJkmgMuAGyJiNvAnsrvpqoAbACLiIqBfSmlo/n5I/tmZwJ8jonGWamlK\nadH6LV2SJG0MKiI0pZSm52syjSc7LfcYcEhK6fW8S19g+yZDhpFdPH5N/mp0Iy0sUyBJktQWFRGa\nAFJK1wLXtvDZySXvP7VeipIkScpVTGiSJGlD1vD229TW1pa7jIrSo0cPqqury11GYYYmSZI6WN1b\nb/Hgo3OZO+nXVFVtVu5yKkavXt0YPvzYThOcDE2SJHWwt5YtY8nS7my66afo1es95S6nItTXL6S2\n9l4aGhoMTZIkaVU9evSkurqlZ9FvfJYuLXcF66bs6zRJkiR1BoYmSZKkAgxNkiRJBRiaJEmSCjA0\nSZIkFWBokiRJKsDQJEmSVIChSZIkqQBDkyRJUgGGJkmSpAIMTZIkSQUYmiRJkgowNEmSJBVgaJIk\nSSrA0CRJklSAoUmSJKkAQ5MkSVIBhiZJkqQCDE2SJEkFGJokSZIKMDRJkiQVYGiSJEkqwNAkSZJU\ngKFJkiSpAEOTJElSAYYmSZKkAgxNkiRJBRiaJEmSCjA0SZIkFWBokiRJKsDQJEmSVIChSZIkqQBD\nkyRJUgGGJkmSpAIMTZIkSQUYmiRJkgowNEmSJBVgaJIkSSrA0CRJklSAoUmSJKkAQ5MkSVIBhiZJ\nkqQCDE2SJEkFGJokSZIKMDRJkiQVYGiSJEkqoGJCU0SMjIi5EbE0Iv4QEfuspf8nI2J2RDRExN8j\nYuj6qnVD8+/X55W7BGmdeMyqM/K47fwqIjRFxHHApcD5wF7A48CMiOjdQv8dgHuAWcCewBXATyLi\n4PVR74bmlQX+Q1bn4jGrzsjjtvOriNAEjAZ+lFK6KaX0N+A0oB74Sgv9TwdeSCmdk1J6NqV0DfDz\nfDuSJEntruyhKSK6AzVks0YApJQSMBP4SAvD9ss/b2rGGvpLkiS1SdlDE9Ab6ArML2mfD/RtYUzf\nFvpvERHvat/yJEmSoFu5C1iPegA8/vLL5a6jovzrjTdYtnwZL774CAsXPl/ucipCQ0Mdb731Eo8/\n/jhbbbVVWWt54403eOnVV3nojTeo7tGjrLVUCo/Z5lXKcesx2zyP29VVyjEL8MwzzzT+dY0HbWRn\nwsonPz1XDxydUrqrSfsNwJYppS80M+b3wOyU0pgmbScBP0wpNfudj4ghwC3tW70kSdqAHJ9SurWl\nD8s+05RSWhYRs4FBwF0AERH5+ytbGPYIcGhJ26fz9pbMAI4HXgQa2lCyJEnasPQAdiDLCi0q+0wT\nQEQcC9xAdtfcn8jugvsiMDCl9HpEXAT0SykNzfvvAPwVuBb4KVnAuhw4LKVUeoG4JElSm5V9pgkg\npTQ9X5NpPNAHeAw4JKX0et6lL7B9k/4vRsRngR8CZwL/BE4xMEmSpI5SETNNkiRJla4SlhyQJEmq\neIamDUz+/L4zy12H1JyIeG9ErIiIPcpdi9QaEXFRRDxc7jpUHoamChER90XEZc20D42IN9ZhUx8C\nJrdfZVJxEfGzPBQ1vhZExG8iYve8y0tk1yg+WcYytZHLj813So7VFU3av7OG4ROAw9ZXraoshqbO\nofCFZyml2pRSuy+pEBFd23ub2mD9huyGjr7AgcBy4G7IHpGUUnotpbSiPXeYr/cmFdUX2C7/cxTw\nJv93zG4HXNLcoIjomlKqTyktbO+CPIY7B0NTJ5L/L/72iPh6RPw7/1/81U0DTZHTcxHx0Yh4NCKW\nRsQfIuJzTU+ZRMQn8vefiYi/REQDsH9EDIiIOyLi1YhYHBF/iohBJdueGxHfiogb8z4v5tvvnY9d\nHBGPR0RNh3yTVAneSim9noejJ4CLge0jolfR03MRMSwiXoqIuoiYHhGjms64RsT5+TF8SkS8ACzN\n2w+JiAcj4o3838fdETGgybjG/R8TEQ9ERH1+HO8UEftExJ/zY/TXEdGrg74/KrP82HwtpfQaWWBK\nTY7Z11JK9fmxtCIiDs6PtbeAmvz03JrWBCQitoiI2yJiSX4cnxoRj0TEhU36vBIR50bELRGxCLgi\nb78sIv6eH5vPRcTYiOjSZNxF+bZOjYiXI2JRRPwwIrrkP3tfzV9f75jv3sbN0NT5fAoYAHwSOBE4\nKX8VEhGbky0i+jiwF3A+8H2an826CDgX2AV4AqgGfpXX8EGyGYW7IuI9JeNGAQ/mfe4BbgZuzP/c\nC3g+f68NXERUAycA/0gp1ebNa5w5jYj9gevIlhT5IHAv8K1mxr0fOAr4Qt4PYDPgUmBvslmud4Db\nm9nNBWRLnOxFNhN2K1m4OwP4WL7t8cW+Sm3gLiT7mbYL8GzetrbZ/2vIjq1DyBZi/mw+vtQ5wB+A\nPcl+DgO8QbYQ80BgDPA1YETJuF2BA4CDyf59jST72bwl2fF7PvCDJqfF1V5SSr4q4AXcB1zWTPtQ\n4D/5338GvEC+VETedhtwa5P3c4Ez17Cf04DXgE2atJ1C9stlj/z9J4AVwOEF6v4rMKJk/zc0ed8n\n39b5Tdo+nO9v23J/33217ys/RpcBi/PXCrJ11D6Yf/7evG2PNWxjKnBXSdvNjf8O8vfnk63sv/Va\n6umd72/Xkv2f1KTPcfnx+IkmbecCT5f7++mr419Nf8aWtB+SHxcHlbRfBDy8hu1tnf8bOLRJWy+y\n2dALm7S9AkwpUN+3gAdK9v8G8K4mbfcBz5SMW+PvAl+teznT1Pk8lfJ/EblXgG2b6xgR1+WnGhbn\n078AOwNPpJTebtL1T80MT8Dsku1tFhGXRMTT+emPxWT/G+pfMvavKzeS0vz8r00v/J0PREt1q9O7\nF9iD7H/P+5A9luC3EbF9aceIeLLJMfqrvPkDrH5MNneMzksp/adke++PiFsj4vmIeJPsF0diDcco\n2fEIqx+jHp+Ckp+DTUXEoKY/YyPiC2SzlF2APzf2S9ks6wtFth0RX46IhxsvgwC+zerH73Mppbea\nvJ/P6jdXeAx3gIpYEVwALCKbWi3Vk+yce6NlJZ8nWj7NOhb4QRtqWlLy/lKyR9Z8newU21LgF8Am\nJf1Kayxtawx9hvYN05KU0tzGNxExjOwYHgZcX9L3UKDxAtil67qfZtruIQtKXwX+TXaMPcWaj9HU\nQpvHp6D546zRQ2T/OWj0KrBba7cdEZ8im639b7L/fCwiu/zilJJxzf0eWJffDWolQ1PleJbs/HSp\nGuDvrdlgSmkBsKCZ/RwfEd1TSo3/yPYtuMmPkp16a3ywcjXZAw5bVV4rx6lzSsCmqzWm9HIzfZ8l\nm6Fqaq3HaERsTTaTekpK6aG87WMt1CK1WcruVF5lBikiniM7BbwP2XWf5DcVDFhtA6v7CPC3lNKl\nTba3Y7sVrDYzhVaO64CdI+LyiNg9InaOiDFk11s0e/trK90KdAV+HBEDI+IQspkjWPWXSTQz9h/A\nURGxZ0TsCdzSQr8iWjtOle9dEdEnfw0ErgKqyG5AKOIq4LCIGJ2fbjsV+AxrDztvALXA8Ih4X0Qc\nSDY7WjquuWPP41HtIj9lPBW4PCIOiIjdyGZY32Ltx/A/gPdHxFGR3a38dVwTqqIYmipEfjrj42TX\nCP0P2R0VXwS+mFL6n3XZ1Fr2sxg4nGxK+VGyhdrG5R83Xd+pue2MIfvF9BBwJ/BbYE6B/Rdt04bh\nM2Snxv5NdhzXkB3HD+afr+0YfZjshoXRZA/v/jTZnXRrXH8sv9bvuHx/fyULTGc317Vgm9RaXyP7\n+fobsp+TM4AXWcvP2JTS/yP7D/SPyH627kl2915reEx3AB/YKyLieLL/CW1ZcnGhVBEi4sfAziml\nT5S7Fmld5Uu9/BsYnlKaWu561Hpe07QRiogTyM7D/4tsfZuLgdsMTKoU+WmJ/yG7UPYwsrVoTi9r\nUVJBEfEhYEeyu+N6ka0LtpTsRgV1YoamjVNfsoX7+pAtWXAb2W2tUqXYF/gGsDlZwD8jpfSz8pYk\nFdYF+CbZ8gNvkS0/8PH88gh1Yp6ekyRJKsALwSVJkgowNEmSJBVgaJIkSSrA0CRJklSAoUmSJKkA\nQ5MkSVIBhiZJkqQCDE2SJEkFGJokbfQiomu5a5BU+QxNkipWRNwXEVdExPciojYiXomI89cy5qMR\n8WhELI2IP0TE5yJiRUTskX/+ifz9ZyLiLxHRAOwfEQMi4o6IeDUiFkfEnyJiUMm250bEtyLixrzP\ni/n2e+djF0fE4xFR04HfFkllYmiSVOlOBOrInkd3DvCd0jDTKH+a/F3A48BewPnA94Hmnhd1EXAu\nsAvwBFAN/Ar4FNmDrH8D3BUR7ykZNwp4MO9zD3AzcGP+517A8/l7SRsYnz0nqWJFxH1Al5TSJ5q0\n/RGYlVI6r5n+p5E9jPo9KaW387ZTgMnAXimlJyLiE8B9wOdTSmt86nxE/BW4LqV0bf5+LvD7lNJJ\n+fvGh16PSymNy9s+DDwMbJdSeq1N3wBJFcWZJkmV7omS968A20bEdfnpsMURsSj/bGfgicbAlPtT\nM9tMwOymDRGxWURcEhFPR8QbEbEYGAj0Lxn715UbSWl+/tcnm3w+Hwhg2yJfnKTOo1u5C5CktVhW\n8j6R/YdvLPCDNmx3Scn7S4FBwNfJTrEtBX4BbLKWekrbGqfv/U+ptIExNEnqlFJKC4AFJc3PAsdH\nRPeUUmOQ2bfgJj8K3JBSugsgIqqBHVpbXivHSapg/k9I0obkVqAr8OOIGBgRh5DNHMGqQSaaGfsP\n4KiI2DMi9gRuaaFfEa0dJ6mCGZokVbJ1mrFJKS0GDgf2BB4FJgDj8o8b1rLdMcAbwEPAncBvgTkF\n6inaJqmT8+45SRu0iDgeuB7YMqX0VrnrkdR5eU2TpA1KRJwAvAD8i2wtpYuB2wxMktrK0CRpQ9OX\nbK2mxjWUbgO+XdaKJG0QPD0nSZJUgBeCS5IkFWBokiRJKsDQJEmSVIChSZIkqQBDkyRJUgGGJkmS\npAIMTZIkSQUYmiRJkgr4/0ez9Dl+o+RUAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2373711fe80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print()\n",
    "start = time.time()\n",
    "results = np.empty(shape=(len(FILES), len(TRANSFORMTYPE), max(len(TRANSFORMINC),len(TRANSFORMDEC)), 5))\n",
    "lambdas_history = []\n",
    "lambdas_history_train = []\n",
    "coverages = []\n",
    "for file in tqdm_notebook(FILES, desc='File', leave = False):\n",
    "    data = get_data(file)\n",
    "    (train_data,heldout_data,test_data) = create_datasets(data)\n",
    "    \n",
    "    train_data = np.concatenate((np.array([\"<<s>>\",\"<s>\"]),train_data))\n",
    "    heldout_data = np.concatenate((np.array([\"<<s>>\",\"<s>\"]),heldout_data))\n",
    "    test_data = np.concatenate((np.array([\"<<s>>\",\"<s>\"]),test_data))\n",
    "    \n",
    "    T = train_data.size\n",
    "    (COUNTS_WORDS,COUNTS_PAIRS,COUNTS_TRIPLES) = create_counts(train_data)\n",
    "    \n",
    "    (lambdas,history) = EM(EPSILON,heldout_data)\n",
    "    lambdas_history.append(history)\n",
    "    \n",
    "    (fold,history) = EM(EPSILON,train_data)\n",
    "    lambdas_history_train.append(history)\n",
    "    \n",
    "    coverages.append(coverage(test_data))\n",
    "    for transform_type in tqdm_notebook(TRANSFORMTYPE, desc='Transf', leave = False):\n",
    "        if transform_type == TRANSFORMTYPE.INCREASE:\n",
    "            for transform_inc in tqdm_notebook(TRANSFORMINC, desc='Inc', leave = False):\n",
    "                new_lambdas = lambda_increase(list(lambdas),transform_inc)\n",
    "                H = entropy(test_data,new_lambdas)\n",
    "                #print(new_lambdas)\n",
    "                #print(H)\n",
    "                results[FILES.index(file),int(transform_type)-1, TRANSFORMINC.index(transform_inc),:] = new_lambdas + [H]\n",
    "                \n",
    "        if transform_type == TRANSFORMTYPE.DECREASE:\n",
    "            for transform_dec in tqdm_notebook(TRANSFORMDEC, desc='Dec', leave = False):\n",
    "                new_lambdas = lambda_decrease(list(lambdas),transform_dec)\n",
    "                H = entropy(test_data,new_lambdas)\n",
    "                #print(new_lambdas)\n",
    "                #print(H)\n",
    "                results[FILES.index(file),int(transform_type)-1, TRANSFORMDEC.index(transform_dec),:] = new_lambdas + [H]\n",
    "\n",
    "                \n",
    "end = time.time()\n",
    "print(\"Time: \",int((end-start)//60) ,\"m \",int((end-start)%60) ,\"s\")\n",
    "print()\n",
    "\n",
    "display(HTML(\"<h1>Průběžné výsledky EM algoritmu</h1>\"))                \n",
    "show_lambdas(lambdas_history)\n",
    "display(HTML(\"<h1>Průběžné výsledky EM algoritmu na trénovacích datech</h1>\"))                \n",
    "show_lambdas(lambdas_history_train)\n",
    "display(HTML(\"<h1>Nastavení $\\lambda_i$ a entropie</h1>\")) \n",
    "show_results(results)\n",
    "display(HTML(\"<h1>Pokrytí</h1>\"))\n",
    "show_coverage(coverages)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Shrnutí výsledků\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Nastavení parametrů a entropie původních textů\n",
    "|             |Anglický text|Český text|\n",
    "|-------------|------------:|---------:|\n",
    "|$\\lambda_0$  |0.070076     |0.140264  |\n",
    "|$\\lambda_1$  |0.254005     |0.428937  |\n",
    "|$\\lambda_2$  |0.492259     |0.244618  |\n",
    "|$\\lambda_3$  |0.183659     |0.186180  |\n",
    "|Entropie     |7.468516     |10.220875 |\n",
    "\n",
    "Vidíme, že v anglickém textu měli největší váhu bi-gramy, zatímco u českého textu to byly uni-gramy.\n",
    "Myslím si, že je to způsobeno tím, že anglický text má daleko vyšší pokrytí a menší slovní zásobu. Angličtina má také pevnou vetnou konstrukci tedy se určité dvojice slov budou vyskytovat často."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Nastavení parametrů při výpočtu na trénovacích datech\n",
    "Při běhu EM algoritmu na trénovacích datech dostaneme 1 u parametru $\\lambda_3$ a u ostatních 0 (na velikosti $\\epsilon$ záleží kolik předchozích parametrů se úplně vynuluje).\n",
    "To je způsobeno tím, že pro všechny trojice dokážeme spočítat pravděpodobnost jejich výskytu a proto, že podmíněná pravděpodobnost trojice (2 slova z historie a jedno aktuální) je větší rovna podmíněné pravděpodobnosti dvojice (pouze jedno slovo z historie). \n",
    "A protože EM algoritmus maximaluzuje očekávanou hodnotu, musí tedy vynulovat všechny parametry kromě $\\lambda_3$."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Zvýšení parametru $\\lambda_3$\n",
    "Při zvyšování parametru $\\lambda_3$ rychle roste cross-entropie. \n",
    "To je dáno tím, že díky zvětšení $\\lambda_3$ máme pocit, že s vyšší pravděpodobností dokážeme odhadnou trojici, ale protože jsme parametr zvětšili uměle, tato představa je milná a nemáme nic, čím bychom to mohli napravit.\n",
    "Tedy čím více zvětšujeme $\\lambda_3$, tím častěji máme chybný odhad a proto se entropie zvětšuje."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Snížení parametru $\\lambda_3$\n",
    "Zde cross-entropie také roste, ale již ne tak výrazně. Je to způsobeno tím, že snížení parametru $\\lambda_3$ nám může ještě \"zachránit\" parametr $\\lambda_2$, sice bychom dosáhli lepšího výsledku s delší historií, ale dvojice jsou podmnožiny trojic, tedy nedojde k takovému zhoršení jako v opačném případě, kdy chceme všechno odhadovat podle trojic."
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [default]",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  },
  "widgets": {
   "state": {
    "19257caf514042419c27d1d04fa3ca66": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    },
    "1af3bb6cedc24dfa9c4f33e583ce3e26": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    },
    "36ed1245da6a4ed9b699986ccae5ddec": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    },
    "3abc32da1d254d3abebef11954e55866": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    },
    "49a7ef51e7b04bbf99ff1922006789e0": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    },
    "4a550a237f474c3d8a7fdbdd31cad892": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    },
    "60f53dec982c43b09256187ab36b9a26": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    },
    "796bdcccc571478d8cb72cf1922c032e": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    },
    "93251f0c5e8748409e937596547573a4": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    },
    "b0b560584a8744f5992a27beeaf0d61e": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    },
    "b99bcbfe83374bae821b21a26c6b7b18": {
     "views": [
      {
       "cell_index": 26
      }
     ]
    }
   },
   "version": "1.2.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}