news_dataset_comparison.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "news dataset comparison.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xddwaic6Uf00",
        "colab_type": "code",
        "outputId": "7df7ead4-f3f1-4dc7-d6db-0b577bf02c46",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "!pip install -U -q PyDrive\n",
        "from pydrive.auth import GoogleAuth\n",
        "from pydrive.drive import GoogleDrive\n",
        "from google.colab import auth\n",
        "from oauth2client.client import GoogleCredentials\n",
        "import warnings\n",
        "\n",
        "# Authenticate and create the PyDrive client.\n",
        "auth.authenticate_user()\n",
        "gauth = GoogleAuth()\n",
        "gauth.credentials = GoogleCredentials.get_application_default()\n",
        "drive = GoogleDrive(gauth)\n",
        "\n",
        "# The dataset shareable link\n",
        "\n",
        "#new stylo data\n",
        "#link = 'https://drive.google.com/open?id=1qg3fx5JyibrfBBycl3y8hCQJWBiXkP2N'\n",
        "link = 'https://drive.google.com/open?id=15e0XWZR1CYx0nyBOLasWJuwMyFgF-U5l'\n",
        "\n",
        "fluff, id = link.split('=')\n",
        "print(id)\n",
        "filename = 'stylo-data.csv'\n",
        "downloaded = drive.CreateFile({'id':id})\n",
        "downloaded.GetContentFile(filename)\n",
        "\n",
        "#news stylo dataset\n",
        "news_link = 'https://drive.google.com/open?id=1C5MS8fWH2yS5GKS_TVddhuXPgR63JC-c'\n",
        "fluff, id = news_link.split('=')\n",
        "print(id)\n",
        "news_filename = 'news-data.csv'\n",
        "downloaded = drive.CreateFile({'id':id})\n",
        "downloaded.GetContentFile(news_filename)\n",
        "\n",
        "# varela stylo dataset\n",
        "varela_link = 'https://drive.google.com/open?id=1_RZDCOie5xAD2IMcrLFFMNZjkzHtZQEl'\n",
        "fluff, id = varela_link.split('=')\n",
        "print(id)\n",
        "varela_filename = 'varela-data.csv'\n",
        "downloaded = drive.CreateFile({'id':id})\n",
        "downloaded.GetContentFile(varela_filename)"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "15e0XWZR1CYx0nyBOLasWJuwMyFgF-U5l\n",
            "1C5MS8fWH2yS5GKS_TVddhuXPgR63JC-c\n",
            "1_RZDCOie5xAD2IMcrLFFMNZjkzHtZQEl\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1zGcn_MPPWv-",
        "colab_type": "code",
        "outputId": "84ed7519-cae7-4008-d065-b8fa0623505f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 258
        }
      },
      "source": [
        "import pandas as pd\n",
        "\n",
        "import numpy as np\n",
        "from matplotlib import pyplot as plt\n",
        "\n",
        "from sklearn.model_selection import train_test_split, cross_validate, StratifiedKFold, RandomizedSearchCV\n",
        "from sklearn.metrics import classification_report, confusion_matrix, accuracy_score, roc_auc_score\n",
        "from sklearn.preprocessing import LabelEncoder, StandardScaler\n",
        "from sklearn.mixture import GaussianMixture, BayesianGaussianMixture\n",
        "from sklearn.cluster import KMeans\n",
        "from sklearn.decomposition import PCA\n",
        "from sklearn.pipeline import Pipeline, make_pipeline\n",
        "from sklearn.linear_model import SGDClassifier, LogisticRegression\n",
        "from sklearn.ensemble import AdaBoostClassifier, BaggingClassifier, ExtraTreesClassifier, GradientBoostingClassifier, RandomForestClassifier\n",
        "from sklearn.svm import SVC, SVR, LinearSVC\n",
        "from sklearn.feature_extraction.text import TfidfTransformer,TfidfVectorizer\n",
        "from sklearn.naive_bayes import GaussianNB, BernoulliNB, MultinomialNB\n",
        "from sklearn.neural_network import MLPClassifier\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "from sklearn.feature_selection import SelectKBest, SelectFromModel, VarianceThreshold, SelectPercentile, SelectFwe, RFECV,  chi2, f_classif, mutual_info_classif\n",
        "from sklearn.preprocessing import StandardScaler, MinMaxScaler, PowerTransformer, QuantileTransformer, Normalizer\n",
        "\n",
        "!pip install -U keras\n",
        "from keras.utils import np_utils\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "from tensorflow.python.keras.layers import InputLayer, Embedding, Conv1D, MaxPool1D, Dense, Flatten, Dropout, GRU, LSTM, Input\n",
        "from tensorflow.python.keras import Sequential, Model\n",
        "from keras.initializers import Constant\n",
        "\n",
        "from pprint import pprint\n",
        "\n",
        "from imblearn.over_sampling import SMOTE"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already up-to-date: keras in /usr/local/lib/python3.6/dist-packages (2.3.1)\n",
            "Requirement already satisfied, skipping upgrade: pyyaml in /usr/local/lib/python3.6/dist-packages (from keras) (3.13)\n",
            "Requirement already satisfied, skipping upgrade: six>=1.9.0 in /usr/local/lib/python3.6/dist-packages (from keras) (1.12.0)\n",
            "Requirement already satisfied, skipping upgrade: keras-applications>=1.0.6 in /usr/local/lib/python3.6/dist-packages (from keras) (1.0.8)\n",
            "Requirement already satisfied, skipping upgrade: numpy>=1.9.1 in /usr/local/lib/python3.6/dist-packages (from keras) (1.18.3)\n",
            "Requirement already satisfied, skipping upgrade: h5py in /usr/local/lib/python3.6/dist-packages (from keras) (2.10.0)\n",
            "Requirement already satisfied, skipping upgrade: keras-preprocessing>=1.0.5 in /usr/local/lib/python3.6/dist-packages (from keras) (1.1.0)\n",
            "Requirement already satisfied, skipping upgrade: scipy>=0.14 in /usr/local/lib/python3.6/dist-packages (from keras) (1.4.1)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/externals/six.py:31: FutureWarning: The module is deprecated in version 0.21 and will be removed in version 0.23 since we've dropped support for Python 2.7. Please rely on the official version of six (https://pypi.org/project/six/).\n",
            "  \"(https://pypi.org/project/six/).\", FutureWarning)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/utils/deprecation.py:144: FutureWarning: The sklearn.neighbors.base module is  deprecated in version 0.22 and will be removed in version 0.24. The corresponding classes / functions should instead be imported from sklearn.neighbors. Anything that cannot be imported from sklearn.neighbors is now part of the private API.\n",
            "  warnings.warn(message, FutureWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "vCg33ap_U4ei",
        "colab": {}
      },
      "source": [
        "# read dataset\n",
        "df = pd.read_csv(filename)\n",
        "df_news = pd.read_csv(news_filename)\n",
        "df_varela = pd.read_csv(varela_filename)\n",
        "\n",
        "# Remove subject column\n",
        "df_varela.pop('Subject')\n",
        "\n",
        "# Remove single class entries (We need at least 3 samples of an author)#\n",
        "df = df.groupby('Author').filter(lambda x: len(x) > 3)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HgSanrNolnvv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Let's drop all the features that are not in our 'best' list\n",
        "\n",
        "best_features = ['DiversidadeLexica', 'TamanhoMedioDasPalavras', 'TamanhoMedioSentencas', 'StdevSentencas', 'TamanhoMedioParagrafos', 'StdevTamParagrafos', 'FrequenciaDeParagrafos',\n",
        "'FrequenciaDeHapaxLegomenaLocal', 'Ponto', 'Virgulas','Exclamacoes', 'DoisPontos', 'FreqCitacoes', 'FreqAdv', \n",
        "'FreqSubs','FreqPrep','FreqVerbosPtcp','FreqPronomes','PronomesPorPreposicao','FreqTermosNaoTageados','FreqPalavrasDeConteudo',\n",
        "'FreqECoisa','FreqEObra','FreqEOutro','FreqEPessoa','FreqETempo','GuiraudR','HerdanV','MedidaK','DugastU','FreqPalavrasComunsTam4',\n",
        "'FreqPalavrasErradas','FrequenciaStopWords','BRFleshIndex']\n",
        "\n",
        "df_news = df_news[best_features]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z0nBZUDPU5mN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Separate data from labels\n",
        "\n",
        "y = df.pop('Author')\n",
        "y_news = df_news.pop('Author')\n",
        "y_varela = df_varela.pop('Author')\n",
        "\n",
        "le = LabelEncoder()\n",
        "le.fit(y)\n",
        "encoded_Y = le.transform(y)\n",
        "\n",
        "le.fit(y_news)\n",
        "encoded_y_news = le.transform(y_news)\n",
        "\n",
        "le = LabelEncoder()\n",
        "le.fit(y_varela)\n",
        "encoded_Y_varela = le.transform(y_varela)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CLqTaPXgU80J",
        "colab_type": "code",
        "outputId": "1a73b290-ecb6-45d8-9096-77ec37d8b2af",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        }
      },
      "source": [
        "# Feature selection\n",
        "\n",
        "# Baseline (Variance threshold):\n",
        "threshold=(.01)\n",
        "selector = VarianceThreshold(threshold)\n",
        "selector.fit(df)\n",
        "\n",
        "#df.columns[selector.get_support(indices=True)] # 20 features, data has very low variance (features are mostly very close)\n",
        "\n",
        "selector.fit(df_news)\n",
        "\n",
        "df_news.columns[selector.get_support(indices=True)]"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['DiversidadeLexica', 'TamanhoMedioDasPalavras', 'TamanhoMedioSentencas',\n",
              "       'StdevSentencas', 'TamanhoMedioParagrafos', 'StdevTamParagrafos',\n",
              "       'MediaSilabasPorPalavra', 'Ponto', 'Virgulas', 'Exclamacoes',\n",
              "       'DoisPontos', 'PronomesPorPreposicao', 'GuiraudR', 'HerdanV', 'MedidaK',\n",
              "       'DugastU', 'BRFleshIndex'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E1hUjRi3U_Si",
        "colab_type": "code",
        "outputId": "ebcfb4cd-2a2b-446d-93dd-6657c3515b17",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 153
        }
      },
      "source": [
        "# 50% better based on percentile using chi2\n",
        "selector = SelectPercentile(chi2, percentile=40)\n",
        "selector.fit(df, y)\n",
        "df.columns[selector.get_support(indices=True)]"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['DiversidadeLexica', 'TamanhoMedioDasPalavras', 'TamanhoMedioSentencas',\n",
              "       'StdevSentencas', 'TamanhoMedioParagrafos', 'StdevTamParagrafos',\n",
              "       'Ponto', 'Virgulas', 'Exclamacoes', 'DoisPontos',\n",
              "       'PronomesPorPreposicao', 'FreqEAcontecimento', 'FreqEOrganizacao',\n",
              "       'FreqEObra', 'FreqEOutro', 'FreqETempo', 'GuiraudR', 'HerdanV',\n",
              "       'DugastU', 'HonoresH', 'FrequenciaDeHapaxLegomenaLocal',\n",
              "       'BRFleshIndex'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F_v7US4sVB_J",
        "colab_type": "code",
        "outputId": "f6a20d3f-0c95-4e61-aad7-6e6d91c64f6c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 170
        }
      },
      "source": [
        "selector = SelectPercentile(chi2, percentile=50)\n",
        "selector.fit(df_news, y_news)\n",
        "df.columns[selector.get_support(indices=True)]"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['DiversidadeLexica', 'TamanhoMedioDasPalavras', 'TamanhoMedioSentencas',\n",
              "       'StdevSentencas', 'TamanhoMedioParagrafos', 'StdevTamParagrafos',\n",
              "       'Ponto', 'Virgulas', 'Exclamacoes', 'DoisPontos', 'FreqCitacoes',\n",
              "       'FreqAdv', 'FreqVerbosPtcp', 'FreqPronomes', 'PronomesPorPreposicao',\n",
              "       'FreqTermosNaoTageados', 'FreqECoisa', 'FreqEObra', 'FreqEOutro',\n",
              "       'FreqEPessoa', 'FreqETempo', 'GuiraudR', 'HerdanV', 'DugastU',\n",
              "       'FreqPalavrasErradas', 'FrequenciaDeHapaxLegomenaLocal',\n",
              "       'BRFleshIndex'],\n",
              "      dtype='object')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4YT-nBFoOCJ7",
        "colab_type": "code",
        "outputId": "20119f2b-47d4-4ef7-9f83-9588516cec56",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "selector.get_support(indices=True)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 0,  1,  2,  3,  4,  5,  9, 10, 11, 12, 13, 15, 20, 22, 23, 24, 31,\n",
              "       34, 35, 36, 37, 39, 41, 43, 46, 47, 52])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cwHSf7U5VEWl",
        "colab_type": "code",
        "outputId": "663e06e2-55ed-41a3-fc64-b09bceea27b6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 952
        }
      },
      "source": [
        "# K-Best (chi2)\n",
        "best_k_features = SelectKBest(chi2, k=54).fit(df, y)\n",
        "dfscores = pd.DataFrame(best_k_features.scores_)\n",
        "dfcolumns = pd.DataFrame(df.columns)\n",
        "featureScores = pd.concat([dfcolumns,dfscores],axis=1)\n",
        "featureScores.columns = ['Feature','Score'] \n",
        "print(featureScores.nlargest(54,'Score'))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                           Feature        Score\n",
            "41                         HerdanV  1640.030726\n",
            "5               StdevTamParagrafos  1190.145679\n",
            "4           TamanhoMedioParagrafos  1083.505879\n",
            "3                   StdevSentencas    83.212074\n",
            "0                DiversidadeLexica    68.963605\n",
            "2            TamanhoMedioSentencas    42.317390\n",
            "43                         DugastU    40.001632\n",
            "10                        Virgulas    12.506046\n",
            "9                            Ponto     8.796392\n",
            "12                      DoisPontos     8.654201\n",
            "39                        GuiraudR     7.570860\n",
            "11                     Exclamacoes     5.398750\n",
            "52                    BRFleshIndex     2.558055\n",
            "23           PronomesPorPreposicao     2.380678\n",
            "47  FrequenciaDeHapaxLegomenaLocal     1.056917\n",
            "1          TamanhoMedioDasPalavras     1.034314\n",
            "45                        HonoresH     0.898729\n",
            "30              FreqEAcontecimento     0.542428\n",
            "37                      FreqETempo     0.538987\n",
            "35                      FreqEOutro     0.433414\n",
            "33                FreqEOrganizacao     0.386386\n",
            "34                       FreqEObra     0.323971\n",
            "6           FrequenciaDeParagrafos     0.316969\n",
            "29                  FreqEAbstracao     0.289026\n",
            "36                     FreqEPessoa     0.276762\n",
            "38                      FreqEValor     0.222563\n",
            "31                      FreqECoisa     0.208482\n",
            "50          FreqPalavrasComunsTam4     0.172185\n",
            "32                      FreqELocal     0.163691\n",
            "7     FrequenciaPalavrasDuplicadas     0.154284\n",
            "42                         MedidaK     0.124415\n",
            "24           FreqTermosNaoTageados     0.119277\n",
            "22                    FreqPronomes     0.105890\n",
            "15                         FreqAdv     0.100361\n",
            "46             FreqPalavrasErradas     0.084512\n",
            "14                   FreqAdjetivos     0.083546\n",
            "13                    FreqCitacoes     0.079495\n",
            "49          FreqPalavrasComunsTam3     0.076480\n",
            "20                  FreqVerbosPtcp     0.074478\n",
            "28      FreqTotalEntidadesNomeadas     0.069403\n",
            "8           MediaSilabasPorPalavra     0.069180\n",
            "21                        FreqConj     0.062580\n",
            "26          FreqPalavrasFuncionais     0.056714\n",
            "48          FreqPalavrasComunsTam2     0.054638\n",
            "16                         FreqArt     0.048959\n",
            "18                        FreqPrep     0.048056\n",
            "53           FreqOperadoresLogicos     0.042504\n",
            "19                        FreqVerb     0.035827\n",
            "51             FrequenciaStopWords     0.034237\n",
            "40                         HerdanC     0.028714\n",
            "27        FrequenciaFrasesNominais     0.025541\n",
            "17                        FreqSubs     0.024090\n",
            "25          FreqPalavrasDeConteudo     0.019389\n",
            "44                           MaasA     0.018078\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nMRUzFOTVJ23",
        "colab_type": "code",
        "outputId": "cde10eba-d74d-4d06-8053-ee05dcc465c3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 952
        }
      },
      "source": [
        "# K-Best (chi2)\n",
        "best_k_features = SelectKBest(chi2, k=54).fit(df_news, y_news)\n",
        "dfscores = pd.DataFrame(best_k_features.scores_)\n",
        "dfcolumns = pd.DataFrame(df.columns)\n",
        "featureScores = pd.concat([dfcolumns,dfscores],axis=1)\n",
        "featureScores.columns = ['Feature','Score'] \n",
        "print(featureScores.nlargest(54,'Score'))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                           Feature        Score\n",
            "41                         HerdanV  2688.503948\n",
            "5               StdevTamParagrafos  1707.831428\n",
            "4           TamanhoMedioParagrafos  1103.327501\n",
            "2            TamanhoMedioSentencas   862.748654\n",
            "11                     Exclamacoes    59.697231\n",
            "3                   StdevSentencas    59.007145\n",
            "0                DiversidadeLexica    37.698913\n",
            "43                         DugastU    33.392413\n",
            "39                        GuiraudR    22.573796\n",
            "10                        Virgulas    21.582665\n",
            "9                            Ponto    16.308883\n",
            "12                      DoisPontos    12.164414\n",
            "23           PronomesPorPreposicao     6.568906\n",
            "52                    BRFleshIndex     4.978543\n",
            "1          TamanhoMedioDasPalavras     1.791226\n",
            "24           FreqTermosNaoTageados     0.810441\n",
            "35                      FreqEOutro     0.551372\n",
            "34                       FreqEObra     0.526174\n",
            "22                    FreqPronomes     0.487441\n",
            "47  FrequenciaDeHapaxLegomenaLocal     0.448172\n",
            "46             FreqPalavrasErradas     0.419544\n",
            "13                    FreqCitacoes     0.399479\n",
            "31                      FreqECoisa     0.389917\n",
            "36                     FreqEPessoa     0.377552\n",
            "37                      FreqETempo     0.349568\n",
            "15                         FreqAdv     0.333384\n",
            "20                  FreqVerbosPtcp     0.316366\n",
            "38                      FreqEValor     0.279046\n",
            "42                         MedidaK     0.263620\n",
            "50          FreqPalavrasComunsTam4     0.252240\n",
            "29                  FreqEAbstracao     0.252198\n",
            "32                      FreqELocal     0.227958\n",
            "17                        FreqSubs     0.222467\n",
            "49          FreqPalavrasComunsTam3     0.205825\n",
            "30              FreqEAcontecimento     0.198315\n",
            "25          FreqPalavrasDeConteudo     0.169387\n",
            "18                        FreqPrep     0.164189\n",
            "6           FrequenciaDeParagrafos     0.163249\n",
            "51             FrequenciaStopWords     0.147728\n",
            "33                FreqEOrganizacao     0.141496\n",
            "8           MediaSilabasPorPalavra     0.141112\n",
            "21                        FreqConj     0.116501\n",
            "53           FreqOperadoresLogicos     0.110885\n",
            "48          FreqPalavrasComunsTam2     0.100544\n",
            "45                        HonoresH     0.099920\n",
            "26          FreqPalavrasFuncionais     0.092239\n",
            "7     FrequenciaPalavrasDuplicadas     0.077802\n",
            "28      FreqTotalEntidadesNomeadas     0.076904\n",
            "16                         FreqArt     0.076476\n",
            "19                        FreqVerb     0.065210\n",
            "27        FrequenciaFrasesNominais     0.060383\n",
            "14                   FreqAdjetivos     0.051196\n",
            "44                           MaasA     0.009959\n",
            "40                         HerdanC     0.006941\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z1JHmkocVEhl",
        "colab_type": "code",
        "outputId": "ecf16734-f5ff-40ce-8b7e-fcc132a5a78b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 544
        }
      },
      "source": [
        "# K-Best (f_classif)\n",
        "best_k_features = SelectKBest(f_classif, k=30).fit(df_news, y_news)\n",
        "dfscores = pd.DataFrame(best_k_features.scores_)\n",
        "dfcolumns = pd.DataFrame(df_news.columns)\n",
        "featureScores = pd.concat([dfcolumns,dfscores],axis=1)\n",
        "featureScores.columns = ['Feature','Score'] \n",
        "print(featureScores.nlargest(30,'Score'))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                           Feature      Score\n",
            "11                     Exclamacoes  19.089614\n",
            "39                        GuiraudR  13.163038\n",
            "22                    FreqPronomes  12.735249\n",
            "25          FreqPalavrasDeConteudo  12.442170\n",
            "24           FreqTermosNaoTageados  12.376033\n",
            "42                         MedidaK  12.154679\n",
            "23           PronomesPorPreposicao  10.844617\n",
            "1          TamanhoMedioDasPalavras  10.182572\n",
            "15                         FreqAdv   9.488145\n",
            "51             FrequenciaStopWords   9.291832\n",
            "41                         HerdanV   9.033462\n",
            "18                        FreqPrep   8.457826\n",
            "17                        FreqSubs   8.408751\n",
            "6           FrequenciaDeParagrafos   8.297498\n",
            "20                  FreqVerbosPtcp   6.958381\n",
            "4           TamanhoMedioParagrafos   6.232975\n",
            "43                         DugastU   6.214138\n",
            "46             FreqPalavrasErradas   6.084080\n",
            "0                DiversidadeLexica   5.951392\n",
            "44                           MaasA   5.548263\n",
            "5               StdevTamParagrafos   5.414068\n",
            "13                    FreqCitacoes   5.060375\n",
            "50          FreqPalavrasComunsTam4   4.877481\n",
            "3                   StdevSentencas   4.812311\n",
            "47  FrequenciaDeHapaxLegomenaLocal   4.667679\n",
            "26          FreqPalavrasFuncionais   4.503462\n",
            "49          FreqPalavrasComunsTam3   4.445219\n",
            "21                        FreqConj   4.396515\n",
            "40                         HerdanC   4.370510\n",
            "10                        Virgulas   4.346741\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I8iXtTWbVVBi",
        "colab_type": "code",
        "outputId": "cec3ffea-338c-41fb-c6cb-6917ed6654dd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 581
        }
      },
      "source": [
        "# Create the RFE object and compute a cross-validated score.\n",
        "rfc = RandomForestClassifier(n_estimators=200)\n",
        "# The \"accuracy\" scoring is proportional to the number of correct classifications\n",
        "\n",
        "rfecv = RFECV(estimator=rfc, step=1, cv=5,scoring='accuracy')\n",
        "rfecv.fit(df_news, y_news)\n",
        "\n",
        "print(\"Optimal number of features news-dataset: %d\" % rfecv.n_features_)\n",
        "\n",
        "rfecv2 = RFECV(estimator=rfc, step=1, cv=3,\n",
        "              scoring='accuracy')\n",
        "rfecv2.fit(df, y)\n",
        "\n",
        "print(\"Optimal number of features students-dataset: %d\" % rfecv2.n_features_)\n",
        "\n",
        "fig, (ax1, ax2) = plt.subplots(1, 2)\n",
        "fig.set_size_inches(12, 8)\n",
        "\n",
        "ax1.plot(range(1, len(rfecv.grid_scores_) + 1), rfecv.grid_scores_)\n",
        "ax1.set_xlabel(\"Number of features selected\")\n",
        "ax1.set_ylabel(\"Cross validation score (nb of correct classifications)\")\n",
        "ax1.set_title('Random Forest n=200 - News dataset')\n",
        "\n",
        "ax2.plot(range(1, len(rfecv2.grid_scores_) + 1), rfecv2.grid_scores_)\n",
        "ax2.set_xlabel(\"Number of features selected\")\n",
        "ax2.set_ylabel(\"Cross validation score (nb of correct classifications)\")\n",
        "ax2.set_title('Random Forest n=200 - Students dataset')\n",
        "\n",
        "plt.figure()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Optimal number of features news-dataset: 40\n",
            "Optimal number of features students-dataset: 49\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4FXajxW9RXU0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X, X_test, Y, y_test = train_test_split(df_news.values, encoded_y_news, random_state=7, test_size=0.3)\n",
        "X_test, X_val, y_test, y_val = train_test_split(X_test, y_test, random_state=7, test_size=0.5)\n",
        "\n",
        "X_student, X_student_test, Y_student, Y_student_test = train_test_split(df.values, encoded_Y, random_state=11, test_size=0.2)\n",
        "X_student_test, X_student_val, Y_student_test, y_student_val = train_test_split(X_student_test, Y_student_test, random_state=11, test_size=0.5)\n",
        "\n",
        "X_varela, X_varela_test, Y_varela, y_varela_test = train_test_split(df_varela, encoded_Y_varela, random_state=7, test_size=0.3)\n",
        "X_varela_test, X_varela_val, y_varela_test, y_varela_val = train_test_split(X_varela_test, y_varela_test, random_state=7, test_size=0.5)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mcDwTcTarJRK",
        "colab_type": "code",
        "outputId": "16c39051-69fc-492a-be9c-b1958918b111",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 224
        }
      },
      "source": [
        "X_varela.head()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>DiversidadeLexica</th>\n",
              "      <th>TamanhoMedioDasPalavras</th>\n",
              "      <th>TamanhoMedioSentencas</th>\n",
              "      <th>StdevSentencas</th>\n",
              "      <th>TamanhoMedioParagrafos</th>\n",
              "      <th>StdevTamParagrafos</th>\n",
              "      <th>FrequenciaDeParagrafos</th>\n",
              "      <th>FrequenciaPalavrasDuplicadas</th>\n",
              "      <th>MediaSilabasPorPalavra</th>\n",
              "      <th>Ponto</th>\n",
              "      <th>Virgulas</th>\n",
              "      <th>Exclamacoes</th>\n",
              "      <th>DoisPontos</th>\n",
              "      <th>FreqCitacoes</th>\n",
              "      <th>FreqAdjetivos</th>\n",
              "      <th>FreqAdv</th>\n",
              "      <th>FreqArt</th>\n",
              "      <th>FreqSubs</th>\n",
              "      <th>FreqPrep</th>\n",
              "      <th>FreqVerb</th>\n",
              "      <th>FreqVerbosPtcp</th>\n",
              "      <th>FreqConj</th>\n",
              "      <th>FreqPronomes</th>\n",
              "      <th>PronomesPorPreposicao</th>\n",
              "      <th>FreqTermosNaoTageados</th>\n",
              "      <th>FreqPalavrasDeConteudo</th>\n",
              "      <th>FreqPalavrasFuncionais</th>\n",
              "      <th>FrequenciaFrasesNominais</th>\n",
              "      <th>FreqTotalEntidadesNomeadas</th>\n",
              "      <th>FreqEAbstracao</th>\n",
              "      <th>FreqEAcontecimento</th>\n",
              "      <th>FreqECoisa</th>\n",
              "      <th>FreqELocal</th>\n",
              "      <th>FreqEOrganizacao</th>\n",
              "      <th>FreqEObra</th>\n",
              "      <th>FreqEOutro</th>\n",
              "      <th>FreqEPessoa</th>\n",
              "      <th>FreqETempo</th>\n",
              "      <th>FreqEValor</th>\n",
              "      <th>GuiraudR</th>\n",
              "      <th>HerdanC</th>\n",
              "      <th>HerdanV</th>\n",
              "      <th>MedidaK</th>\n",
              "      <th>DugastU</th>\n",
              "      <th>MaasA</th>\n",
              "      <th>HonoresH</th>\n",
              "      <th>FreqPalavrasErradas</th>\n",
              "      <th>FrequenciaDeHapaxLegomenaLocal</th>\n",
              "      <th>FreqPalavrasComunsTam2</th>\n",
              "      <th>FreqPalavrasComunsTam3</th>\n",
              "      <th>FreqPalavrasComunsTam4</th>\n",
              "      <th>FrequenciaStopWords</th>\n",
              "      <th>BRFleshIndex</th>\n",
              "      <th>FreqOperadoresLogicos</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1283</th>\n",
              "      <td>77.631579</td>\n",
              "      <td>5.389831</td>\n",
              "      <td>23.000000</td>\n",
              "      <td>7.071068</td>\n",
              "      <td>69.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.013158</td>\n",
              "      <td>0.026316</td>\n",
              "      <td>2.065789</td>\n",
              "      <td>3.947368</td>\n",
              "      <td>2.631579</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.026316</td>\n",
              "      <td>0.052632</td>\n",
              "      <td>0.039474</td>\n",
              "      <td>0.105263</td>\n",
              "      <td>0.171053</td>\n",
              "      <td>0.157895</td>\n",
              "      <td>0.013158</td>\n",
              "      <td>0.092105</td>\n",
              "      <td>0.039474</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.197368</td>\n",
              "      <td>0.394737</td>\n",
              "      <td>0.302632</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.092105</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.142857</td>\n",
              "      <td>0.285714</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.285714</td>\n",
              "      <td>0.285714</td>\n",
              "      <td>6.767764</td>\n",
              "      <td>0.941535</td>\n",
              "      <td>59.0</td>\n",
              "      <td>2.225427</td>\n",
              "      <td>106.866298</td>\n",
              "      <td>0.096734</td>\n",
              "      <td>0.864407</td>\n",
              "      <td>0.105263</td>\n",
              "      <td>0.671053</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.394737</td>\n",
              "      <td>423.520658</td>\n",
              "      <td>0.039474</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1888</th>\n",
              "      <td>53.703704</td>\n",
              "      <td>6.800000</td>\n",
              "      <td>26.333333</td>\n",
              "      <td>18.168349</td>\n",
              "      <td>395.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.002315</td>\n",
              "      <td>0.046296</td>\n",
              "      <td>2.243056</td>\n",
              "      <td>3.472222</td>\n",
              "      <td>4.861111</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.064815</td>\n",
              "      <td>0.048611</td>\n",
              "      <td>0.081019</td>\n",
              "      <td>0.196759</td>\n",
              "      <td>0.115741</td>\n",
              "      <td>0.020833</td>\n",
              "      <td>0.134259</td>\n",
              "      <td>0.067130</td>\n",
              "      <td>0.062500</td>\n",
              "      <td>0.540000</td>\n",
              "      <td>0.069444</td>\n",
              "      <td>0.495370</td>\n",
              "      <td>0.263889</td>\n",
              "      <td>0.013889</td>\n",
              "      <td>0.111111</td>\n",
              "      <td>0.020833</td>\n",
              "      <td>0.104167</td>\n",
              "      <td>0.041667</td>\n",
              "      <td>0.104167</td>\n",
              "      <td>0.104167</td>\n",
              "      <td>0.041667</td>\n",
              "      <td>0.125000</td>\n",
              "      <td>0.062500</td>\n",
              "      <td>0.291667</td>\n",
              "      <td>0.104167</td>\n",
              "      <td>11.162105</td>\n",
              "      <td>0.897554</td>\n",
              "      <td>232.0</td>\n",
              "      <td>2.510466</td>\n",
              "      <td>85.458244</td>\n",
              "      <td>0.108174</td>\n",
              "      <td>0.801724</td>\n",
              "      <td>0.050926</td>\n",
              "      <td>0.430556</td>\n",
              "      <td>0.006944</td>\n",
              "      <td>0.011574</td>\n",
              "      <td>0.013889</td>\n",
              "      <td>0.416667</td>\n",
              "      <td>438.545810</td>\n",
              "      <td>0.074074</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1454</th>\n",
              "      <td>61.320755</td>\n",
              "      <td>5.776923</td>\n",
              "      <td>23.250000</td>\n",
              "      <td>14.838716</td>\n",
              "      <td>186.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.004717</td>\n",
              "      <td>0.056604</td>\n",
              "      <td>1.995283</td>\n",
              "      <td>3.773585</td>\n",
              "      <td>6.132075</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.471698</td>\n",
              "      <td>0.009434</td>\n",
              "      <td>0.042453</td>\n",
              "      <td>0.033019</td>\n",
              "      <td>0.099057</td>\n",
              "      <td>0.169811</td>\n",
              "      <td>0.117925</td>\n",
              "      <td>0.009434</td>\n",
              "      <td>0.108491</td>\n",
              "      <td>0.033019</td>\n",
              "      <td>0.028302</td>\n",
              "      <td>0.240000</td>\n",
              "      <td>0.141509</td>\n",
              "      <td>0.457547</td>\n",
              "      <td>0.250000</td>\n",
              "      <td>0.018868</td>\n",
              "      <td>0.113208</td>\n",
              "      <td>0.208333</td>\n",
              "      <td>0.041667</td>\n",
              "      <td>0.083333</td>\n",
              "      <td>0.125000</td>\n",
              "      <td>0.125000</td>\n",
              "      <td>0.083333</td>\n",
              "      <td>0.125000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.125000</td>\n",
              "      <td>0.083333</td>\n",
              "      <td>8.928437</td>\n",
              "      <td>0.908701</td>\n",
              "      <td>130.0</td>\n",
              "      <td>2.380399</td>\n",
              "      <td>84.643939</td>\n",
              "      <td>0.108693</td>\n",
              "      <td>0.800000</td>\n",
              "      <td>0.070755</td>\n",
              "      <td>0.490566</td>\n",
              "      <td>0.004717</td>\n",
              "      <td>0.009434</td>\n",
              "      <td>0.009434</td>\n",
              "      <td>0.344340</td>\n",
              "      <td>417.597642</td>\n",
              "      <td>0.033019</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1523</th>\n",
              "      <td>44.954955</td>\n",
              "      <td>6.494990</td>\n",
              "      <td>17.627451</td>\n",
              "      <td>11.263268</td>\n",
              "      <td>899.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000901</td>\n",
              "      <td>0.053153</td>\n",
              "      <td>1.995495</td>\n",
              "      <td>3.873874</td>\n",
              "      <td>8.738739</td>\n",
              "      <td>0.270270</td>\n",
              "      <td>0.810811</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.059459</td>\n",
              "      <td>0.045946</td>\n",
              "      <td>0.090090</td>\n",
              "      <td>0.137838</td>\n",
              "      <td>0.079279</td>\n",
              "      <td>0.009910</td>\n",
              "      <td>0.091892</td>\n",
              "      <td>0.045946</td>\n",
              "      <td>0.065766</td>\n",
              "      <td>0.829545</td>\n",
              "      <td>0.127027</td>\n",
              "      <td>0.381982</td>\n",
              "      <td>0.215315</td>\n",
              "      <td>0.010811</td>\n",
              "      <td>0.091892</td>\n",
              "      <td>0.058824</td>\n",
              "      <td>0.137255</td>\n",
              "      <td>0.058824</td>\n",
              "      <td>0.088235</td>\n",
              "      <td>0.098039</td>\n",
              "      <td>0.098039</td>\n",
              "      <td>0.049020</td>\n",
              "      <td>0.009804</td>\n",
              "      <td>0.245098</td>\n",
              "      <td>0.156863</td>\n",
              "      <td>14.977491</td>\n",
              "      <td>0.885982</td>\n",
              "      <td>499.0</td>\n",
              "      <td>2.684614</td>\n",
              "      <td>88.725646</td>\n",
              "      <td>0.106164</td>\n",
              "      <td>0.779559</td>\n",
              "      <td>0.084685</td>\n",
              "      <td>0.350450</td>\n",
              "      <td>0.014414</td>\n",
              "      <td>0.018018</td>\n",
              "      <td>0.018018</td>\n",
              "      <td>0.352252</td>\n",
              "      <td>417.608198</td>\n",
              "      <td>0.037838</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1696</th>\n",
              "      <td>41.788478</td>\n",
              "      <td>6.374486</td>\n",
              "      <td>24.631579</td>\n",
              "      <td>17.504016</td>\n",
              "      <td>936.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000860</td>\n",
              "      <td>0.081685</td>\n",
              "      <td>1.964746</td>\n",
              "      <td>2.837489</td>\n",
              "      <td>9.114359</td>\n",
              "      <td>0.171969</td>\n",
              "      <td>0.687876</td>\n",
              "      <td>0.049871</td>\n",
              "      <td>0.030095</td>\n",
              "      <td>0.042132</td>\n",
              "      <td>0.084265</td>\n",
              "      <td>0.161651</td>\n",
              "      <td>0.076526</td>\n",
              "      <td>0.020636</td>\n",
              "      <td>0.105761</td>\n",
              "      <td>0.045572</td>\n",
              "      <td>0.061909</td>\n",
              "      <td>0.808989</td>\n",
              "      <td>0.124678</td>\n",
              "      <td>0.437661</td>\n",
              "      <td>0.206363</td>\n",
              "      <td>0.012898</td>\n",
              "      <td>0.085985</td>\n",
              "      <td>0.070000</td>\n",
              "      <td>0.080000</td>\n",
              "      <td>0.030000</td>\n",
              "      <td>0.170000</td>\n",
              "      <td>0.060000</td>\n",
              "      <td>0.030000</td>\n",
              "      <td>0.090000</td>\n",
              "      <td>0.080000</td>\n",
              "      <td>0.310000</td>\n",
              "      <td>0.080000</td>\n",
              "      <td>14.251035</td>\n",
              "      <td>0.876388</td>\n",
              "      <td>486.0</td>\n",
              "      <td>2.665571</td>\n",
              "      <td>82.383655</td>\n",
              "      <td>0.110174</td>\n",
              "      <td>0.707819</td>\n",
              "      <td>0.049011</td>\n",
              "      <td>0.295787</td>\n",
              "      <td>0.017197</td>\n",
              "      <td>0.017197</td>\n",
              "      <td>0.017197</td>\n",
              "      <td>0.351677</td>\n",
              "      <td>415.014140</td>\n",
              "      <td>0.036113</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      DiversidadeLexica  ...  FreqOperadoresLogicos\n",
              "1283          77.631579  ...               0.039474\n",
              "1888          53.703704  ...               0.074074\n",
              "1454          61.320755  ...               0.033019\n",
              "1523          44.954955  ...               0.037838\n",
              "1696          41.788478  ...               0.036113\n",
              "\n",
              "[5 rows x 54 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MvjESmURRgbI",
        "colab_type": "code",
        "outputId": "4f41bacc-f815-48ad-ddbb-dd687c904e63",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "model = ExtraTreesClassifier(n_estimators=350, random_state=11)\n",
        "\n",
        "model.fit(X, Y)\n",
        "\n",
        "y_pred = model.predict(X_test)\n",
        "\n",
        "score = accuracy_score(y_test, y_pred)\n",
        "print(\"News Test score: %.2f%%\" % (score * 100))\n",
        "\n",
        "\n",
        "model = ExtraTreesClassifier(n_estimators=350, random_state=11)\n",
        "model.fit(X_student, Y_student)\n",
        "\n",
        "y_pred = model.predict(X_student_test)\n",
        "\n",
        "score = accuracy_score(Y_student_test, y_pred)\n",
        "print(\"Students Test score: %.2f%%\" % (score * 100))\n",
        "\n",
        "\n",
        "model = ExtraTreesClassifier(n_estimators=800, random_state=11)\n",
        "model.fit(X_varela, Y_varela)\n",
        "\n",
        "y_pred = model.predict(X_varela_test)\n",
        "\n",
        "score = accuracy_score(y_varela_test, y_pred)\n",
        "print(\"Varela Test score: %.2f%%\" % (score * 100))"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "News Test score: 73.33%\n",
            "Students Test score: 14.29%\n",
            "Varela Test score: 56.67%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1iA5Gg74j5Nx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pep2ZehISvJK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Hyper-tuning our classifiers\n",
        "\n",
        "# TODO: Extract this into reusable function!\n",
        "\n",
        "n_estimators = [int(x) for x in np.linspace(start = 150, stop = 1500, num = 50)]\n",
        "max_features = ['auto', 'sqrt']\n",
        "max_depth = [int(x) for x in np.linspace(10, 110, num = 11)]\n",
        "max_depth.append(None)\n",
        "min_samples_split = [2, 5, 10]\n",
        "min_samples_leaf = [1, 2, 4]\n",
        "bootstrap = [True, False]\n",
        "\n",
        "random_grid = {'n_estimators': n_estimators,\n",
        "               'max_features': max_features,\n",
        "               'max_depth': max_depth,\n",
        "               'min_samples_split': min_samples_split,\n",
        "               'min_samples_leaf': min_samples_leaf,\n",
        "               'bootstrap': bootstrap}\n",
        "                \n",
        "etc = ExtraTreesClassifier(random_state=11)\n",
        "\n",
        "# search across 100 different combinations, and use all available cores\n",
        "news_best_estimator_search = RandomizedSearchCV(estimator = etc, param_distributions = random_grid, n_iter = 100, cv = 5, verbose=0, random_state=11, n_jobs=-1)\n",
        "\n",
        "news_best_estimator_search.fit(X, Y)\n",
        "\n",
        "students_best_estimator_search = RandomizedSearchCV(estimator = etc, param_distributions = random_grid, n_iter = 100, cv = 3, verbose=0, random_state=11, n_jobs=-1)\n",
        "\n",
        "students_best_estimator_search.fit(X_student, Y_student)\n",
        "\n",
        "\n",
        "pprint(news_best_estimator_search.best_params_)\n",
        "pprint(students_best_estimator_search.best_params_)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Eb7EeVT7KyEb",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "ff563e7e-5861-41ef-da47-f0e134382cf7"
      },
      "source": [
        "Y_varela.shape"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2100,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jl8FGC0HKwBJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 173
        },
        "outputId": "86a53c3e-f819-4b0d-90f4-ef55e1c0e5cc"
      },
      "source": [
        "varela_best_estimator_search = RandomizedSearchCV(estimator = etc, param_distributions = random_grid, n_iter = 100, cv = 5, verbose=0, random_state=11, n_jobs=-1)\n",
        "\n",
        "varela_best_estimator_search.fit(X_varela, Y_varela)\n",
        "pprint(varela_best_estimator_search.best_params_)"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/joblib/externals/loky/process_executor.py:706: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
            "  \"timeout or by a memory leak.\", UserWarning\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "{'bootstrap': False,\n",
            " 'max_depth': 60,\n",
            " 'max_features': 'sqrt',\n",
            " 'min_samples_leaf': 1,\n",
            " 'min_samples_split': 2,\n",
            " 'n_estimators': 866}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BGPCA-Jnf5e2",
        "colab_type": "code",
        "outputId": "90c3e517-1289-4d1b-f46d-c035dcbd39b9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "from sklearn.metrics import average_precision_score\n",
        "from sklearn.metrics import precision_recall_fscore_support\n",
        "\n",
        "# Let's now check the accuracy on both validation datasets (not seen by the classifier yet)\n",
        "new_X = np.concatenate((X, X_test))\n",
        "new_Y = np.concatenate((Y, y_test))\n",
        "\n",
        "news_best_estimator_search.best_estimator_.fit(new_X, new_Y)\n",
        "predictions = news_best_estimator_search.best_estimator_.predict(X_val)\n",
        "base_score = accuracy_score(predictions, y_val)\n",
        "print('Average accuracy news:' + str(base_score))\n",
        "#pprint(classification_report(y_val, predictions))\n",
        "\n",
        "new_X_student = np.concatenate((X_student, X_student_test))\n",
        "new_Y_student = np.concatenate((Y_student, Y_student_test))\n",
        "\n",
        "students_best_estimator_search.best_estimator_.fit(new_X_student, new_Y_student)\n",
        "predictions = students_best_estimator_search.best_estimator_.predict(X_student_val)\n",
        "base_score = accuracy_score(predictions, y_student_val)\n",
        "print('Average accuracy students:' + str(base_score))\n",
        "\n",
        "new_X_varela = np.concatenate((X_varela, X_varela_test))\n",
        "new_Y_varela = np.concatenate((Y_varela, y_varela_test))\n",
        "\n",
        "varela_best_estimator_search.best_estimator_.fit(new_X_varela, new_Y_varela)\n",
        "predictions = varela_best_estimator_search.best_estimator_.predict(X_varela_val)\n",
        "varela_score = accuracy_score(predictions, y_varela_val)\n",
        "print('Accuracy varela:' + str(varela_score))\n"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Average accuracy news:0.8\n",
            "Average accuracy students:0.14285714285714285\n",
            "Accuracy varela:0.5644444444444444\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_UV62Xl8RCWR",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        },
        "outputId": "4188c709-7aee-4ef8-c81e-6186e4cdfb0f"
      },
      "source": [
        "rfecv_varela = RFECV(estimator=ExtraTreesClassifier(n_estimators=866), step=1, cv=5,scoring='accuracy')\n",
        "rfecv_varela.fit(df_varela, encoded_Y_varela)\n",
        "\n",
        "varela_best_features_num = rfecv_varela.n_features_\n",
        "\n",
        "print(\"Optimal number of features for best estimator news-dataset: %d\" % varela_best_features_num)\n",
        "\n",
        "fig, (ax1) = plt.subplots(1, 1)\n",
        "fig.set_size_inches(12, 8)\n",
        "\n",
        "ax1.plot(range(1, len(rfecv_varela.grid_scores_) + 1), rfecv_varela.grid_scores_)\n",
        "ax1.set_xlabel(\"Number of features selected\")\n",
        "ax1.set_ylabel(\"Cross validation score (nb of correct classifications)\")\n",
        "ax1.set_title('Best classifier (ExtraTrees) - Varela dataset')\n",
        "\n",
        "\n",
        "plt.figure()"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Optimal number of features for best estimator news-dataset: 54\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7LPhbq7lIreZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "# print the metrics if needed\n",
        "\n",
        "# precision_recall_fscore_support(y_test, predictions, average='macro')\n",
        "\n",
        "# pprint(classification_report(y_student_val, predictions))\n",
        "\n",
        " #print (\"AUC&ROC\",roc_auc_score(y_test, predictions, multi_class='ovr'))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GIGbPqTN25-V",
        "colab_type": "code",
        "outputId": "6457b24b-96b0-486c-ee65-0a5f573cf7d9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 581
        }
      },
      "source": [
        "rfecv = RFECV(estimator=news_best_estimator_search.best_estimator_, step=1, cv=5,scoring='accuracy')\n",
        "rfecv.fit(df_news, encoded_y_news)\n",
        "\n",
        "news_best_features_num = rfecv.n_features_\n",
        "\n",
        "print(\"Optimal number of features for best estimator news-dataset: %d\" % news_best_features_num)\n",
        "\n",
        "rfecv2 = RFECV(estimator=students_best_estimator_search.best_estimator_, step=1, cv=3,\n",
        "              scoring='accuracy')\n",
        "rfecv2.fit(df, encoded_Y)\n",
        "\n",
        "students_best_features_num = rfecv2.n_features_\n",
        "\n",
        "print(\"Optimal number of features for best estimator students-dataset: %d\" % students_best_features_num)\n",
        "\n",
        "fig, (ax1, ax2) = plt.subplots(1, 2)\n",
        "fig.set_size_inches(12, 8)\n",
        "\n",
        "ax1.plot(range(1, len(rfecv.grid_scores_) + 1), rfecv.grid_scores_)\n",
        "ax1.set_xlabel(\"Number of features selected\")\n",
        "ax1.set_ylabel(\"Cross validation score (nb of correct classifications)\")\n",
        "ax1.set_title('Best classifier (ExtraTrees) - News dataset')\n",
        "\n",
        "ax2.plot(range(1, len(rfecv2.grid_scores_) + 1), rfecv2.grid_scores_)\n",
        "ax2.set_xlabel(\"Number of features selected\")\n",
        "ax2.set_ylabel(\"Cross validation score (nb of correct classifications)\")\n",
        "ax2.set_title('Best classifier (ExtraTrees) - Students dataset')\n",
        "\n",
        "plt.figure()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Optimal number of features for best estimator news-dataset: 22\n",
            "Optimal number of features for best estimator students-dataset: 26\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 99
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jN_DZ_XO7K1Z",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Dummy function to make label names shorter and help visualize our graph\n",
        "\n",
        "def make_label(text):\n",
        "  if len(text) > 6:\n",
        "    sub_text = text[:6]\n",
        "    capitals = ''.join([c for c in text if c.isupper()])\n",
        "    numbers = ''.join([c for c in text if c.isdigit()])\n",
        "    text = capitals[0] + sub_text[1:] + capitals[1:] + numbers \n",
        "  return text"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5N4gO4Ipu1ah",
        "colab_type": "code",
        "outputId": "014ea275-f4f0-40ed-9b28-49f4a8b529d3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 278
        }
      },
      "source": [
        "fig, (ax1, ax2) = plt.subplots(1, 2)\n",
        "fig.set_size_inches(60, 8)\n",
        "\n",
        "news_features = news_feat_importances.nlargest(news_best_features_num)\n",
        "height_news = [i for i in news_features]\n",
        " \n",
        "ax1.set_xlabel(\"Feature name\")\n",
        "ax1.set_ylabel(\"Feature score\")\n",
        "ax1.set_title('News dataset best features')\n",
        "smaller_news_labels = [make_label(ft) for ft in news_features.index]\n",
        "ax1.bar(smaller_news_labels, height_news, 0.5)\n",
        "\n",
        "std_features = student_feat_importances.nlargest(students_best_features_num)\n",
        "height_std = [i for i in std_features]\n",
        "ax2.set_xlabel(\"Feature name\")\n",
        "ax2.set_ylabel(\"Feature score\")\n",
        "ax2.set_title('News dataset best features')\n",
        "smaller_std_labels = [make_label(ft) for ft in std_features.index]\n",
        "ax2.bar(smaller_std_labels, height_std, 0.5)\n",
        "\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAADVMAAAHwCAYAAAAR2R7zAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdf7Bmd10f8PcneyFBlKSGBWIScqNJB0MY23FJrIVCxWhwxYUhkQia2KIRmVgda+u1HSNmsLOxY5k6jdXYAHEtTTAOurqhEYcC1WKaDQQhiRmXsE42AckvAgEDrPn0j3u283Bzd3M3yfM8e+95vWaeued8z/d8v59z7n/Pd97Pt7o7AAAAAAAAAAAAAAAAABvdUfMuAAAAAAAAAAAAAAAAAGAWhKkAAAAAAAAAAAAAAACAURCmAgAAAAAAAAAAAAAAAEZBmAoAAAAAAAAAAAAAAAAYBWEqAAAAAAAAAAAAAAAAYBSEqQAAAAAAAAAAAAAAAIBREKYCAAD+v6p6eVXtm3cdh6uqPlBVPzaHeZ9bVR+qqi9U1a/Nen4AAAAAAABYD6xDHva81iEBAGCKhKkAAGCGqmpvVX22qp450fZjVfWBOZb1hFTVj1bVn22UeZ7gvBcnuS/Js7r7Xz/J+d5ZVW99MmMAAAAAAAAwTtYhj9x5nuC81iEBAGCKhKkAAGD2NiX56XkXwVPilCS3dXfPu5CqWph3DQAAAAAAAMyVdciNwzokAABMkTAVAADM3n9M8nNVddxqF6vqBVX1vqp6oKruqKofHNpPrarPVdVRw/lvV9VnJ+7bUVU/Mxz/aFXdWVVfqKpPVdUbDjLXM4ZfInuwqm5L8uIV15eq6pPDOLdV1WuG9m9N8ptJ/klVPVxVnxvat1bVR6vq81V1V1W9ZWKsY6rqd6vq/uE5bqqq5w7Xjq2qq6rq01V1d1W9tao2HWyeg/iWqvq/w9x/WFXfODH3d1TV/xnm/VhVvXzi2mPe1Vrmrap3Jrkoyb8d+nx3VR018c7ur6p3r6jj96rqM1X1UFV9qKpeOLRfnOQNE2P90dDeVXXa5JwHfjWuql5eVfuq6uer6jNJ3nGo+Q/1/gEAAAAAANgQrENah7QOCQAAayBMBQAAs7c7yQeS/NzKC1X1zCTvS/KuJM9JckGS36iqM7r7U0k+n+QfD93/WZKHhy/ck+RlST44jPHrSV7Z3d+Q5DuT3HKQWn4pybcMn+/N8pfykz6Z5KVJjk3yy0l+t6pO6O7bk7wpyYe7++u7+8CCzBeTXJjkuCRbk/xkVb16uHbRMM7JSY4f7v+74do7k+xPctrwfN+T5McOMc9qLkzyL5OcMIz160lSVScm2ZXkrUm+Mcvv/feravPB3tVa5u3uH03y35P86tDnT5P8VJJXZ/l/8U1JHkxyxcRt701yepb/tx8Z7k93X7lirFcd4jknPW94plOSXPw48x/q/QMAAAAAALD+WYe0DmkdEgAA1kCYCgAA5uPSJD9VVZtXtH9/kr3d/Y7u3t/dH03y+0nOH65/MMnLqup5w/l1w/mpSZ6V5GND+6NJzqyqZ3T3p7v71oPU8YNJfqW7H+juuzJ88X9Ad/9ed9/T3Y9297VJ/jrJWQd7qO7+QHd/fOj/l0n+R5a/UE+Sr2b5y/PTuvvvu/vm7v788Ktk35fkZ7r7i9392SRvy/ICzuHY0d2f6O4vJvnFJD9YVZuS/HCS67v7+qGu92V5Ien7DvNdrcWbkvz77t7X3V9O8pYk51XVQpJ099u7+wsT176tqo59EvM9muSXuvvL3f13jzP/qu//ScwNAAAAAADAkcc6pHXIt8Q6JAAAHJIwFQAAzEF3fyLJHydZWnHplCRnV9XnDnySvCHLv/yVLC9ivDzLvwb3oSz/stzLhs//Hr6k/2KS12X5C+1PV9WuqnrBQUr5piR3TZz/zeTFqrqwqm6ZqOXMJM8+2HNV1dlV9b+q6t6qemio4UD/HUluSHJNVd1TVb9aVU8bnvlpQ60H5vmtLP9q2uFY+RxPG+Y+Jcn5K97pS5KccJjvai1OSfKeiXluT/L3SZ5bVZuqantVfbKqPp9k73DPQd/nGtzb3Y+sZf4c/P0DAAAAAACwQViHtA4Z65AAAPC4hKkAAGB+finJjyc5caLtriQf7O7jJj5f390/OVz/YJKXZnkh44NJ/izJP83yIsYHDwzS3Td09zlJTkjyV0l++yA1fDrJyRPnzz9wUFWnDPddkuT47j4uySeS1IFpVhnvXUl2Jjm5u49N8psH+nf3V7v7l7v7jCTfmeVfv7tweOYvJ3n2xDM/q7tfeIh5VrPyOb6a5L5h/B0r3ukzu3v7UNfB3tVa5510V5JXrpjrmO6+O8nrk2xL8t1Jjk2yONxzqPf5pSRfN3H+vBXXV95z0PkP8f4BAAAAAADYWKxDWodMrEMCAMBBCVMBAMCcdPeeJNcm+VcTzX+c5B9W1Y9U1dOGz4ur6luHe/46yd8l+eEsL3Z8PsnfJnlthkWMqnpuVW2rqmdmeXHg4SSPHqSMdyf5har6B1V1UpKfmrj2zCx/SX7vMO6/yPIvwh3wt0lOqqqnT7R9Q5IHuvuRqjory1/cZ7j/n1fVi6pqU5LPZ3mR4dHu/nSSP0nya1X1rKo6qqq+papedoh5VvPDVXVGVX1dksuSXNfdf5/kd5O8qqq+d/hVtmOq6uVVddLjvKu1zjvpN5P8yrAAlKraXFXbJt7Nl5Pcn+WFif+w4t6/TfLNK9puSfL6oe5zs7xY9YTmP9j7P4xnAwAAAAAAYB2wDmkdcsW91iEBAGAFYSoAAJivy7K8WJAk6e4vJPmeJBckuSfJZ5JcnuToiXs+mOT+7r5r4rySfGQ4PyrJzw73P5DlL75/Mqv75SR/k+RTWV5I2DFRy21Jfi3Jh7P8BfuLkvz5xL3vT3Jrks9U1X1D25uTXFZVX0hyaZYXSQ54XpLrsvwF+u1D3QfmuzDJ05PcluTBod8Jh5hnNTuSvDPL7+yYDItDw3valuTfZXlB5q4k/ybL7+lQ72qt8076z1n+Rbw/Gd7BXyQ5e7j2O1l+13cPz/kXK+69KskZVfW5qvqDoe2nk7wqyeeSvCHJH+TQDjX/od4/AAAAAAAAG4t1yGXWIa1DAgDAY1T3E9kxFgAAAAAAAAAAAAAAAGB9sTMVAAAAAAAAAAAAAAAAMArCVAAAAAAAAAAAAAAAAMAoCFMBAAAAAAAAAAAAAAAAoyBMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAAAAAAAAAAAwCgvzLmAWnv3sZ/fi4uK8ywAAAAAAAFZx880339fdm+ddB8Dhsg4JAAAAAABHroOtQ44iTLW4uJjdu3fPuwwAAAAAAGAVVfU3864B4ImwDgkAAAAAAEeug61DHjXrQgAAAAAAAAAAAAAAAADmQZgKAAAAAAAAAAAAAAAAGAVhKgAAAAAAAAAAAAAAAGAUhKkAAAAAAAAAAAAAAACAURCmAgAAAAAAAAAAAAAAAEZBmAoAAAAAAAAAAAAAAAAYBWEqAAAAAAAAAAAAAAAAYBSEqQAAAAAAAAAAAAAAAIBREKYCAAAAAAAAAAAAAAAARkGYCgAAAAAAAAAAAAAAABgFYSoAAAAAAAAAAAAAAABgFISpAAAAAAAAAAAAAAAAgFEQpgIAAAAAAAAAAAAAAABGQZgKAAAAAAAAAAAAAAAAGAVhKgAAAAAAAAAAAAAAAGAUhKkAAAAAAAAAAAAAAACAURCmAgAAAAAAAAAAAAAAAEZBmAoAAAAAAAAAAAAAAAAYBWEqAAAAAAAAAAAAAAAAYBQW5l0A87G4tGveJSRJ9m7fOu8SAAAAAAAAgKfILNchrTUCAAAAAPBE2JkKAAAAAAAAAAAAAAAAGAVhKgAAAAAAAGBdqapzq+qOqtpTVUurXD+6qq4drt9YVYtD+1lVdcvw+VhVvWatYwIAAAAAABvDwrwLAAAAAAAAAFirqtqU5Iok5yTZl+SmqtrZ3bdNdHtjkge7+7SquiDJ5Ulel+QTSbZ09/6qOiHJx6rqj5L0GsZkxBaXds1srr3bt85sLgAAAACAMbIzFQAAAAAAALCenJVkT3ff2d1fSXJNkm0r+mxLcvVwfF2SV1RVdfeXunv/0H5MlkNUax0TAAAAAADYAISpAAAAAAAAgPXkxCR3TZzvG9pW7TOEpx5KcnySVNXZVXVrko8nedNwfS1jZrj/4qraXVW777333qfgcQAAAAAAgFkSpgIAAAAAAABGo7tv7O4XJnlxkl+oqmMO8/4ru3tLd2/ZvHnzdIoEAAAAAACmRpgKAAAAAAAAWE/uTnLyxPlJQ9uqfapqIcmxSe6f7NDdtyd5OMmZaxwTAAAAAADYAISpAAAAAAAAgPXkpiSnV9WpVfX0JBck2bmiz84kFw3H5yV5f3f3cM9CklTVKUlekGTvGscEAAAAAAA2gIV5FwDztri0a94lJEn2bt867xIAAAAAAACOeN29v6ouSXJDkk1J3t7dt1bVZUl2d/fOJFcl2VFVe5I8kOVwVJK8JMlSVX01yaNJ3tzd9yXJamPO9MEAAAAAAICZEKYCAAAAAAAA1pXuvj7J9SvaLp04fiTJ+avctyPJjrWOCQAAAAAAbDxHTXPwqjq3qu6oqj1VtbTK9aOr6trh+o1Vtbji+vOr6uGq+rm1jgkAAAAAAAAAAAAAAACwmqmFqapqU5IrkrwyyRlJfqiqzljR7Y1JHuzu05K8LcnlK67/pyTvPcwxAQAAAAAAAAAAAAAAAB5jmjtTnZVkT3ff2d1fSXJNkm0r+mxLcvVwfF2SV1RVJUlVvTrJp5LcephjAgAAAAAAAAAAAAAAADzGNMNUJya5a+J839C2ap/u3p/koSTHV9XXJ/n5JL/8BMYEAAAAAAAAAAAAAAAAeIxphqmejLckeVt3P/xEB6iqi6tqd1Xtvvfee5+6ygAAAAAAAAAAAAAAAIB1aWGKY9+d5OSJ85OGttX67KuqhSTHJrk/ydlJzquqX01yXJJHq+qRJDevYcwkSXdfmeTKJNmyZUs/6acBAAAAAAAAAAAAAAAA1rVphqluSnJ6VZ2a5cDTBUlev6LPziQXJflwkvOSvL+7O8lLD3Soqrckebi7/8sQuHq8MQEAAAAAAAAAAAAAAAAeY2phqu7eX1WXJLkhyaYkb+/uW6vqsiS7u3tnkquS7KiqPUkeyHI46rDHnNYzAAAAAAAAAAAAAAAAABvHNHemSndfn+T6FW2XThw/kuT8xxnjLY83JgAAAAAAAAAAAAAAAMDjOWreBQAAAAAAAAAAAAAAAADMgjAVAAAAAAAAAAAAAAAAMArCVAAAAAAAAAAAAAAAAMAoCFMBAAAAAAAAAAAAAAAAoyBMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAAAAAAAAAAAwCsJUAAAAAAAAAAAAAAAAwCgIUwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjIIwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKAhTAQAAAAAAAAAAAAAAAKMgTAUAAAAAAAAAAAAAAACMwsK8CwAAAAAAAAAA5mNxadfM5tq7fevM5gIAAAAAOBg7UwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjIIwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKCzMuwAAAAAAAAAAgKfS4tKumc21d/vWmc0FAAAAADx5dqYCAAAAAAAAAAAAAAAARkGYCgAAAAAAAAAAAAAAABgFYSoAAAAAAAAAAAAAAABgFISpAAAAAAAAAAAAAAAAgFEQpgIAAAAAAAAAAAAAAABGQZgKAAAAAAAAAAAAAAAAGIWFeRcAAAAAAAAAAMDjW1zaNbO59m7fOrO5AAAAAGCW7EwFAAAAAAAAAAAAAAAAjIIwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKAhTAQAAAAAAAAAAAAAAAKMgTAUAAAAAAAAAAAAAAACMgjAVAAAAAAAAAAAAAAAAMArCVAAAAAAAAAAAAAAAAMAoCFMBAAAAAAAAAAAAAAAAoyBMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAAAAAAAAAAAwCsJUAAAAAAAAAAAAAAAAwCgIUwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjMLCvAsAAAAAAAAAAGC8Fpd2zWyuvdu3zmwuAAAAAI5MdqYCAAAAAAAAAAAAAAAARmGqYaqqOreq7qiqPVW1tMr1o6vq2uH6jVW1OLSfVVW3DJ+PVdVrJu7ZW1UfH67tnmb9AAAAAAAAAAAAAAAAwMaxMK2Bq2pTkiuSnJNkX5Kbqmpnd9820e2NSR7s7tOq6oIklyd5XZJPJNnS3fur6oQkH6uqP+ru/cN9/7y775tW7QAAAAAAAAAAAAAAAMDGM82dqc5Ksqe77+zuryS5Jsm2FX22Jbl6OL4uySuqqrr7SxPBqWOS9BTrBAAAAAAAAAAAAAAAAEZgmmGqE5PcNXG+b2hbtc8QnnooyfFJUlVnV9WtST6e5E0T4apO8idVdXNVXTzF+gEAAAAAAAAAAAAAAIANZGHeBRxMd9+Y5IVV9a1Jrq6q93b3I0le0t13V9Vzkryvqv6quz+08v4haHVxkjz/+c+fae0AAAAAAAAAAAAAAADAkWeaO1PdneTkifOThrZV+1TVQpJjk9w/2aG7b0/ycJIzh/O7h7+fTfKeJGetNnl3X9ndW7p7y+bNm5/0wwAAAAAAAAAAAAAAAADr2zTDVDclOb2qTq2qpye5IMnOFX12JrloOD4vyfu7u4d7FpKkqk5J8oIke6vqmVX1DUP7M5N8T5JPTPEZAAAAAAAAAAAAAAAAgA1iYVoDd/f+qrokyQ1JNiV5e3ffWlWXJdnd3TuTXJVkR1XtSfJAlgNXSfKSJEtV9dUkjyZ5c3ffV1XfnOQ9VXWg9nd19/+c1jMAAAAAAAAAAAAAAAAAG8fUwlRJ0t3XJ7l+RdulE8ePJDl/lft2JNmxSvudSb7tqa8UAAAAAAAAAAAAAAAA2OiOmncBAAAAAAAAAIejqs6tqjuqak9VLa1y/eiquna4fmNVLQ7t51TVzVX18eHvd03c84FhzFuGz3Nm90QAAAAAAMCsTHVnKgAAAAAAAICnUlVtSnJFknOS7EtyU1Xt7O7bJrq9McmD3X1aVV2Q5PIkr0tyX5JXdfc9VXVmkhuSnDhx3xu6e/dMHgQAAAAAAJgLO1MBAAAAAAAA68lZSfZ0953d/ZUk1yTZtqLPtiRXD8fXJXlFVVV3f7S77xnab03yjKo6eiZVAwAAAAAARwRhKgAAAAAAAGA9OTHJXRPn+/K1u0t9TZ/u3p/koSTHr+jz2iQf6e4vT7S9o6puqapfrKpabfKquriqdlfV7nvvvffJPAcAAAAAADAHwlQAAAAAAADAqFTVC5NcnuQnJprf0N0vSvLS4fMjq93b3Vd295bu3rJ58+bpFwsAAAAAADylhKkAAAAAAACA9eTuJCdPnJ80tK3ap6oWkhyb5P7h/KQk70lyYXd/8sAN3X338PcLSd6V5Kwp1Q8AAAAAAMzRwrwLAAAAAAAAADgMNyU5vapOzXJo6oIkr1/RZ2eSi5J8OMl5Sd7f3V1VxyXZlWSpu//8QOchcHVcd99XVU9L8v1J/nT6jwLARre4tGtmc+3dvnVmcwEAAACsZ3amAgAAAAAAANaN7t6f5JIkNyS5Pcm7u/vWqrqsqn5g6HZVkuOrak+Sn02yNLRfkuS0JJdW1S3D5zlJjk5yQ1X9ZZJbshzS+u3ZPRUAAAAAADArdqYCAAAAAAAA1pXuvj7J9SvaLp04fiTJ+avc99Ykbz3IsN/+VNYIAAAAAAAcmexMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAAAAAAAAAAAwCsJUAAAAAAAAAAAAAAAAwCgIUwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjIIwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKAhTAQAAAAAAAAAAAAAAAKOwMO8CAAAAAAAAAACA9WNxadfM5tq7fevM5gIAAADGwc5UAAAAAAAAAAAAAAAAwCgIUwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjIIwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKAhTAQAAAAAAAAAAAAAAAKOwMO8CAAAAAAAAAAAA5m1xadfM5tq7fevM5gIAAAC+lp2pAAAAAAAAAAAAAAAAgFGwMxUAAAAAAAAAAMAGtZF33NrIzwYAAMD0CFMBAAAAAAAAAADAEUJIDAAAYLqEqQAAAAAAAAAAAICpEhIDAACOFMJUwNeY5ZcWh+ILDQAAAAAAAAAAAAAA4Kl21LwLAAAAAAAAAAAAAAAAAJgFYSoAAAAAAAAAAAAAAABgFBbmXQDAkWpxade8S0iS7N2+dd4lAAAAAAAAAAAAAADAhmBnKgAAAAAAAAAAAAAAAGAUhKkAAAAAAAAAAAAAAACAUViYdwEAAAAAAAAAAAAA69Hi0q6ZzbV3+9aZzQUAABuZnakAAAAAAAAAAAAAAACAURCmAgAAAAAAAAAAAAAAAEZBmAoAAAAAAAAAAAAAAAAYBWEqAAAAAAAAAAAAAAAAYBSEqQAAAAAAAAAAAAAAAIBRmGqYqqrOrao7qmpPVS2tcv3oqrp2uH5jVS0O7WdV1S3D52NV9Zq1jgkAAAAAAAAAAAAAAACwmqmFqapqU5IrkrwyyRlJfqiqzljR7Y1JHuzu05K8LcnlQ/snkmzp7n+U5Nwkv1VVC2scEwAAAAAAAAAAAAAAAOAxprkz1VlJ9nT3nd39lSTXJNm2os+2JFcPx9cleUVVVXd/qbv3D+3HJOnDGBMAAAAAAAAAAAAAAADgMRamOPaJSe6aON+X5OyD9enu/VX1UJLjk9xXVWcneXuSU5L8yHB9LWMCAAAAAAAAAAAA8CQsLu2a2Vx7t2+d2Vwb9bkAAFi7ae5M9aR0943d/cIkL07yC1V1zOHcX1UXV9Xuqtp97733TqdIAAAAAAAAAAAAAAAAYN2YZpjq7iQnT5yfNLSt2qeqFpIcm+T+yQ7dfXuSh5OcucYxD9x3ZXdv6e4tmzdvfhKPAQAAAAAAAAAAAAAAAGwE0wxT3ZTk9Ko6taqenuSCJDtX9NmZ5KLh+Lwk7+/uHu5ZSJKqOiXJC5LsXeOYAAAAAAAAAAAAAAAAAI+xMK2Bu3t/VV2S5IYkm5K8vbtvrarLkuzu7p1Jrkqyo6r2JHkgy+GoJHlJkqWq+mqSR5O8ubvvS5LVxpzWMwAAAAAAAAAAAAAAAAAbx9TCVEnS3dcnuX5F26UTx48kOX+V+3Yk2bHWMQEAAAAAAAAAAAAAAAAez1TDVAAAAAAAAAAAAADAdC0u7ZrZXHu3b53ZXAAA0yBMBQAAAAAAAAAAAAAccTZySGwjPxsAHOmEqQAAAAAAAAAAAAAAeNKExABYD4SpAAAAAAAAAAAAAADgIITEADaWo+ZdAAAAAAAAAAAAAAAAAMAs2JkKgDWZ5a8qHIpfXAAAAAAAAAAAAAAA4ImyMxUAAAAAAAAAAAAAAAAwCsJUAAAAAAAAAAAAAAAAwCgIUwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjIIwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKCzMuwAAWG8Wl3bNu4Qkyd7tW+ddAgAAAAAAAAAAAADAumJnKgAAAAAAAAAAAAAAAGAUhKkAAAAAAAAAAAAAAACAUViYdwEAwPq2uLRr3iUkSfZu3zrvEgAAAAAAAAAAAACAI5wwFQAAAAAAAAAAAAAAjMysf0zdj6YDRwphKgAAAAAAAAAAAAAAYMOYZVBMSAzWn6PmXQAAAAAAAADA4aiqc6vqjqraU1VLq1w/uqquHa7fWFWLQ/s5VXVzVX18+PtdE/d8+9C+p6p+vapqdk8EAAAAAADMijAVAAAAAAAAsG5U1aYkVyR5ZZIzkvxQVZ2xotsbkzzY3acleVuSy4f2+5K8qrtflOSiJDsm7vmvSX48yenD59ypPQQAAAAAADA3wlQAAAAAAADAenJWkj3dfWd3fyXJNUm2reizLcnVw/F1SV5RVdXdH+3ue4b2W5M8Y9jF6oQkz+ruv+juTvI7SV49/UcBAAAAAABmbWHeBQAAAAAAAAAchhOT3DVxvi/J2Qfr0937q+qhJMdneWeqA16b5CPd/eWqOnEYZ3LME1ebvKouTnJxkjz/+c9/Eo8BAAAAAHB4Fpd2zWyuvdu3zmwumDU7UwEAAAAAAACjUlUvTHJ5kp843Hu7+8ru3tLdWzZv3vzUFwcAAAAAAEyVnakAAAAAAACA9eTuJCdPnJ80tK3WZ19VLSQ5Nsn9SVJVJyV5T5ILu/uTE/1PepwxAQAAAACYAjtuMWt2pgIAAAAAAADWk5uSnF5Vp1bV05NckGTnij47k1w0HJ+X5P3d3VV1XJJdSZa6+88PdO7uTyf5fFV9R1VVkguT/OG0HwQAAAAAAJg9YSoAAAAAAABg3eju/UkuSXJDktuTvLu7b62qy6rqB4ZuVyU5vqr2JPnZJEtD+yVJTktyaVXdMnyeM1x7c5L/lmRPkk8mee9snggAAAAAAJilhXkXAAAAAAAAAHA4uvv6JNevaLt04viRJOevct9bk7z1IGPuTnLmU1spAAAAAABwpLEzFQAAAAAAAAAAAAAAADAKdqYCAHiKLC7tmncJSZK927fOuwQAAAAAAAAAAAAAOCLZmQoAAAAAAAAAAAAAAAAYBWEqAAAAAAAAAAAAAAAAYBSEqQAAAAAAAAAAAAAAAIBRWJh3AQAAAAAAAAAAAAAAALCRLC7tmul8e7dvnel865mdqQAAAAAAAAAAAAAAAIBREKYCAAAAAAAAAAAAAAAARkGYCgAAAAAAAAAAAAAAABgFYSoAAAAAAAAAAAAAAABgFISpAAAAAAAAAAAAAAAAgFFYmHcBAABsTItLu+ZdQpJk7/at8y7BuwAAAAAAAAAAAAA4QtiZCnS0vtIAACAASURBVAAAAAAAAAAAAAAAABgFYSoAAAAAAAAAAAAAAABgFKYapqqqc6vqjqraU1VLq1w/uqquHa7fWFWLQ/s5VXVzVX18+PtdE/d8YBjzluHznGk+AwAAAAAAAAAAAAAAALAxLExr4KralOSKJOck2Zfkpqra2d23TXR7Y5IHu/u0qrogyeVJXpfkviSv6u57qurMJDckOXHivjd09+5p1Q4AAAAAAAAAAAAAAABsPGvemaqqvu4wxz4ryZ7uvrO7v5LkmiTbVvTZluTq4fi6JK+oquruj3b3PUP7rUmeUVVHH+b8AAAAAAAAwBHuCaxDAgAAAAAAPGGPG6aqqu+sqtuS/NVw/m1V9RtrGPvEJHdNnO/L1+4u9TV9unt/koeSHL+iz2uTfKS7vzzR9o6quqWqfrGq6iB1X1xVu6tq9/9j745j/TrL+4B/H/uSgBgJrTFdGyc4XVymJBVV47lsK50qREgVWkdbsrilEKkR6QRZV7FOclclohFIyaa1a0eEFAgliYQCysq4WgwpKLQdrKQ2bTZw2kwmuI29VjUhDZA1UKfP/vBxdXV7r+/NtX+/n+89n490dN/znve89zl/P/r+3mPHjq2iXAAAAAAAAGBaTqMPCQAAAAAAsGarOZnqV5O8MclTSdLd/yvJj0yyqJOq6rIkdyT52QXTb+7u70/yuuF6y1Lvdvdd3b2zu3du3bp18sUCAAAAAAAAL8TM+pAAAAAAAMB4rSZMle5+ctHU86t47WiSCxfcbxvmllxTVXNJzs/QLKmqbUk+luSt3f3lBbUcHf5+I8mHk+xazTcAAAAAAAAAZ5c19iEBAAAAAADWbDVhqier6p8k6ap6UVX9QpI/WsV7+5PsqKqLq+qcJHuSzC9aM5/khmF8bZKHu7ur6uVJHkyyt7s/d3JxVc1V1SuG8YuSvCnJl1ZRCwAAAAAAAHB2WWsfEgAAAAAAYM1WE6b6V0nekeSCnDhJ6geG+1Pq7uNJbk7yUE40PT7a3Qer6raq+olh2d1JtlTVoSTvTLJ3mL85ySVJbq2qR4frlUnOTfJQVf3vJI8O9bx/dZ8KAAAAAAAAnEXW1IcEAAAAAAA4HXOnelhVm5P8Wne/eS2bd/e+JPsWzd26YPxckuuWeO/dSd69zLZXrKUWAAAAAAAA4Oxwun1IAAAAAACAtTrlyVTd/XySV1XVOVOqBwAAAAAAANjg9CEBAAAAAIBZOeXJVIMnknyuquaTPHtysrt/ZWJVAQAAG9b2vQ/OuoQcvv3qWZcAAAAA6EMCAAAAAAAzsJow1ZeHa1OSl022HAAAAAAAAGAk9CEBAAAAAICpWzFM1d2/nCRV9feG+29OuigAAIAxOBtO6Uqc1AUAAMBs6EMCAAAAAACzsGmlBVV1eVX9YZKDSQ5W1Req6rLJlwYAAAAAAABsVPqQAAAAAADALKwYpkpyV5J3dveruvtVSf5tkvdPtiwAAAAAAABgg9OHBAAAAAAApm41YaqXdvdnTt50928neenEKgIAAAAAAADGQB8SAAAAAACYurlVrHmiqm5Jct9w/9NJnphcSQAAAIzR9r0PzrqEHL796lmXAAAAMCb6kAAAAAAAwNStJkz1M0l+OclvJukk/2OYAwAAACbgbAiWJcJlAADAxOlDAgAAAAAAU7dimKq7n07yc1OoBQAAAAAAABgJfUgAAAAAAGAWVgxTVdWnklzX3X853H9Hkvu7+42TLg4AAADgbDipyyldAABw5ulDAgAAAAAAs7BpFWtecbKBkfztL8S9cnIlAQAAAAAAACOgDwkAAAAAAEzdiidTJfmbqrqou/80SarqVUl6smUBAAAAsNjZcEpX4qQuAADOGH1IAAAAAABg6lYTpvqlJJ+tqt9JUklel+SmiVYFAAAAAAAAbHT6kAAAAAAAwNStGKbq7k9W1Q8mee0w9fPd/dXJlgUAAAAAAABsZPqQAAAAAADALGxaaUFV/dMkf9Xd/z3Jy5P8+6p61cQrAwAAAAAAADYsfUgAAAAAAGAWVgxTJXlfkv9XVa9J8s4kX05y70SrAgAAAAAAADY6fUgAAAAAAGDqVhOmOt7dnWR3kju7+84kL5tsWQAAAAAAAMAGpw8JAAAAAABM3dwq1nyjqn4xyU8n+ZGq2pTkRZMtCwAAAAAAANjg9CEBAAAAAICpW83JVNcn+VaSG7v7z5NsS/IfJ1oVAAAAAAAAsNHpQwIAAAAAAFO34slUQ+PiVxbc/2mSeydZFAAAAAAAALCx6UMCAAAAAACzsJqTqQAAAAAAAAAAAAAAAADWvRVPpgIAAACAs9H2vQ/OuoQcvv3qWZcAAAAAAAAAAMALsKqTqarqJVX16kkXAwAAAAAAAIyHPiQAAAAAADBtK4apqurHkzya5JPD/Q9U1fykCwMAAAAAAAA2Ln1IAAAAAABgFlZzMtW7kuxK8pdJ0t2PJrl4gjUBAAAAAAAAG9+7og8JAAAAAABM2WrCVH/d3c8smutJFAMAAAAAAACMhj4kAAAAAAAwdXOrWHOwqn4qyeaq2pHk55L8z8mWBQAAAAAAAGxw+pAAAAAAAMDUreZkqn+d5LIk30ry4STPJPn5SRYFAAAAAAAAbHj6kAAAAAAAwNSd8mSqqtqc5MHu/tEkvzSdkgAAAAAAAICNTB8SAAAAAACYlVOeTNXdzyf5m6o6f0r1AAAAAAAAABucPiQAAAAAADArpzyZavDNJF+sqk8lefbkZHf/3MSqAgAAAAAAADY6fUgAAAAAAGDqVhOm+s3hAgAAAAAAADhT9CEBAAAAAICpWzFM1d33TKMQAAAAAAAAYDz0IQEAAAAAgFlYMUxVVV9J0ovnu/t7J1IRAAAAAAAAsOHpQwIAAAAAALOwYpgqyc4F4xcnuS7Jd06mHAAAAAAAAGAk9CEBAAAAAICpWzFM1d1PLZr6z1X1hSS3TqYkAAAAAOCF2L73wVmXkCQ5fPvVsy4BAFhHTqcPWVVXJfm1JJuTfKC7b1/0/Nwk9ya5IslTSa7v7sNVtSXJA0n+UZIPdffNC9757STfneSvhqkru/sv1vJtAAAAAADA2WvFMFVV/eCC20058QtxqznRCgAAAAAAAGBJa+1DVtXmJHcmeUOSI0n2V9V8dz+2YNmNSZ7u7kuqak+SO5Jcn+S5JLckuXy4Fntzdx9Yy/cAAAAAAADrw2pCUf9pwfh4kq8k+ZeTKQcAAAAAAAAYibX2IXclOdTdTyRJVd2fZHeShWGq3UneNYwfSPLeqqrufjbJZ6vqktOsHQAAAAAAWKdWE6a68WQj4qSqunhC9QAAAAAAnJbtex+cdQk5fPvVsy4BANaDtfYhL0jy5IL7I0l+aLk13X28qp5JsiXJV1fY+zeq6vkk/zXJu7u7V1EPAAAAAACwjmxaxZoHVjkHAAAAAAAAsFpnWx/yzd39/UleN1xvWWpRVd1UVQeq6sCxY8emWiAAAAAAAHD6lj2Zqqr+YZLLkpxfVf98waPzkrx4NZtX1VVJfi3J5iQf6O7bFz0/N8m9Sa5I8lSS67v7cFW9IcntSc5J8u0k/667Hx7euSLJh5K8JMm+JP/GL8IBAAAAAADA+nAG+pBHk1y44H7bMLfUmiNVNZfk/JzoRy6ru48Of79RVR9OsisnepmL192V5K4k2blzpz4lAAAAAACsM8uGqZK8Osmbkrw8yY8vmP9GkrettHFVbU5yZ5I3JDmSZH9VzXf3YwuW3Zjk6e6+pKr2JLkjyfVJvprkx7v7/1bV5UkeSnLB8M77hv//SE6Eqa5K8omV6gEAAAAAAADOCqfVh0yyP8mOqro4J0JTe5L81KI180luSPJ7Sa5N8vCpfqBxCFy9vLu/WlUvGur79Oo+BwAAAAAAWE+WDVN198eTfLyq/nF3/94a9t6V5FB3P5EkVXV/kt1JFoapdid51zB+IMl7q6q6+w8XrDmY5CXDKVbfmeS87v78sOe9Sa6JMBUAAAAAAACsC6fbh+zu41V1c078IOPmJB/s7oNVdVuSA909n+TuJPdV1aEkX8uJwFWSpKoO58QpWOdU1TVJrkzyJ0keGoJUm3MiSPX+0/lOAAAAAADg7HSqk6lO+sOqekeSy5K8+ORkd//MCu9dkOTJBfdHkvzQcmuGpsczSbbkxMlUJ/2LJH/Q3d+qqguGfRbueUGWUFU3JbkpSS666KIVSgUAAAAA2Hi2731w1iUkSQ7ffvWsSwDg7LTWPmS6e1+SfYvmbl0wfi7Jdcu8u32Zba9YuWQAAAAAAGC927SKNfcl+ftJ3pjkd5JsS/KNSRZ1UlVdluSOJD/7Qt/t7ru6e2d379y6deuZLw4AAAAAAAA4HTPrQwIAAAAAAOO1mjDVJd19S5Jnu/ueJFfn754wtZSjSS5ccL9tmFtyTVXNJTk/yVPD/bYkH0vy1u7+8oL121bYEwAAAAAAADj7rbUPCQAAAAAAsGZzq1jz18Pfv6yqy5P8eZJXruK9/Ul2VNXFORF42pPkpxatmU9yQ5LfS3Jtkoe7u6vq5UkeTLK3uz93cnF3/1lVfb2qXpvkkSRvTfJfVlELAAAAAAAjt33vg7MuIYdvv3rWJQCcTdbahwQAAAAAAFiz1ZxMdVdVfUeSW3Ii/PRYkv+w0kvdfTzJzUkeSvJHST7a3Qer6raq+olh2d1JtlTVoSTvTLJ3mL85ySVJbq2qR4frZOPk7Uk+kORQki8n+cQqvgEAAAAAAAA4u6ypDwkAAAAAAHA6VjyZqrs/MAx/J8n3vpDNu3tfkn2L5m5dMH4uyXVLvPfuJO9eZs8DSS5/IXUAAAAAAAAAZ5fT6UMCAAAAAACs1YonU1XVd1XV3VX1ieH+0qq6cfKlAQAAAAAAABuVPiQAAAAAADALK4apknwoyUNJvme4/z9Jfn5SBQEAAAAAAACj8KHoQwIAAAAAAFO2mjDVK7r7o0n+Jkm6+3iS5ydaFQAAAAAAALDR6UMCAAAAAABTt5ow1bNVtSVJJ0lVvTbJMxOtCgAAAAAAANjo9CEBAAAAAICpm1vFmncmmU/yD6rqc0m2Jrl2olUBAAAAAAAAG50+JAAAAAAAMHXLhqmq6qLu/tPu/oOq+mdJXp2kkjze3X89tQoBAAAAAACADUMfEgAAAAAAmKVNp3j23xaMP9LdB7v7SxoYAAAAAAAAwGnQhwQAAAAAAGbmVGGqWjD+3kkXAgAAAAAAAIyCPiQAAAAAADAzpwpT9TJjAAAAAAAAgLXShwQAAAAAAGZm7hTPXlNVX8+JX4Z7yTDOcN/dfd7EqwMAAAAAAAA2Gn1IAAAAAABgZpYNU3X35mkWAgAAAAAAAGx8+pAAAAAAAMAsbZp1AQAAAAAAAAAAAAAAAADTIEwFAAAAAAAAAAAAAAAAjIIwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKAhTAQAAAAAAAAAAAAAAAKMgTAUAAAAAAAAAAAAAAACMgjAVAAAAAAAAAAAAAAAAMArCVAAAAAAAAAAAAAAAAMAoCFMBAAAAAAAAAAAAAAAAoyBMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAAAAAAAAAAAwCsJUAAAAAAAAAAAAAAAAwCgIUwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjIIwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKAhTAQAAAAAAAAAAAAAAAKMgTAUAAAAAAAAAAAAAAACMgjAVAAAAAAAAAAAAAAAAMArCVAAAAAAAAAAAAAAAAMAoCFMBAAAAAAAAAAAAAAAAoyBMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAAAAAAAAAAAwCsJUAAAAAAAAAAAAAAAAwCgIUwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjIIwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKAhTAQAAAAAAAAAAAAAAAKMw0TBVVV1VVY9X1aGq2rvE83Or6iPD80eqavswv6WqPlNV36yq9y5657eHPR8drldO8hsAAAAAAAAAAAAAAACAjWFuUhtX1eYkdyZ5Q5IjSfZX1Xx3P7Zg2Y1Jnu7uS6pqT5I7klyf5LkktyS5fLgWe3N3H5hU7QAAAAAAAAAAAAAAAMDGM8mTqXYlOdTdT3T3t5Pcn2T3ojW7k9wzjB9I8vqqqu5+trs/mxOhKgAAAAAAAAAAAAAAAIDTNskw1QVJnlxwf2SYW3JNdx9P8kySLavY+zeq6tGquqWq6kwUCwAAAAAAAAAAAAAAAGxskwxTTcqbu/v7k7xuuN6y1KKquqmqDlTVgWPHjk21QAAAAAAAAAAAAAAAAODsM8kw1dEkFy643zbMLbmmquaSnJ/kqVNt2t1Hh7/fSPLhJLuWWXdXd+/s7p1bt25d0wcAAAAAAAAAAAAAAAAAG8ckw1T7k+yoqour6pwke5LML1ozn+SGYXxtkoe7u5fbsKrmquoVw/hFSd6U5EtnvHIAAAAAAAAAAAAAAABgw5mb1Mbdfbyqbk7yUJLNST7Y3Qer6rYkB7p7PsndSe6rqkNJvpYTgaskSVUdTnJeknOq6pokVyb5kyQPDUGqzUk+neT9k/oGAAAAAAAAAAAAAAAAYOOYWJgqSbp7X5J9i+ZuXTB+Lsl1y7y7fZltrzhT9QEAAAAAAAAAAAAAAADjsWnWBQAAAAAAAAAAAAAAAABMgzAVAAAAAAAAAAAAAAAAMArCVAAAAAAAAAAAAAAAAMAoCFMBAAAAAAAAAAAAAAAAoyBMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAACwrlTVVVX1eFUdqqq9Szw/t6o+Mjx/pKq2D/NbquozVfXNqnrvoneuqKovDu/8elXVdL4GAAAAAACYJmEqAAAAAAAAYN2oqs1J7kzyY0kuTfKTVXXpomU3Jnm6uy9J8qtJ7hjmn0tyS5JfWGLr9yV5W5Idw3XVma8eAAAAAACYNWEqAAAAAAAAYD3ZleRQdz/R3d9Ocn+S3YvW7E5yzzB+IMnrq6q6+9nu/mxOhKr+VlV9d5Lzuvvz3d1J7k1yzUS/AgAAAAAAmAlhKgAAAAAAAGA9uSDJkwvujwxzS67p7uNJnkmyZYU9j6ywJwAAAAAAsAEIUwEAAAAAAACsUlXdVFUHqurAsWPHZl0OAAAAAADwAglTAQAAAAAAAOvJ0SQXLrjfNswtuaaq5pKcn+SpFfbctsKeSZLuvqu7d3b3zq1bt77A0gEAAAAAgFkTpgIAAAAAAADWk/1JdlTVxVV1TpI9SeYXrZlPcsMwvjbJw93dy23Y3X+W5OtV9dqqqiRvTfLxM186AAAAAAAwa3OzLgAAAAAAAABgtbr7eFXdnOShJJuTfLC7D1bVbUkOdPd8kruT3FdVh5J8LScCV0mSqjqc5Lwk51TVNUmu7O7Hkrw9yYeSvCTJJ4YLAAAAAADYYISpAAAAAAAAgHWlu/cl2bdo7tYF4+eSXLfMu9uXmT+Q5PIzVyUAAAAAAHA22jTrAgAAAAAAAAAAAAAAAACmQZgKAAAAAAAAAAAAAAAAGAVhKgAAAAAAAAAAAAAAAGAUhKkAAAAAAAAAAAAAAACAURCmAgAAAAAAAAAAAAAAAEZBmAoAAAAAAAAAAAAAAAAYBWEqAAAAAAAAAAAAAAAAYBSEqQAAAAAAAAAAAAAAAIBREKYCAAAAAAAAAAAAAAAARkGYCgAAAAAAAAAAAAAAABgFYSoAAAAAAAAAAAAAAABgFISpAAAAAAAAAAAAAAAAgFEQpgIAAAAAAAAAAAAAAABGQZgKAAAAAAAAAAAAAAAAGAVhKgAAAAAAAAAAAAAAAGAUhKkAAAAAAAAAAAAAAACAURCmAgAAAAAAAAAAAAAAAEZBmAoAAAAAAAAAAAAAAAAYBWEqAAAAAAAAAAAAAAAAYBSEqQAAAAAAAAAAAAAAAIBREKYCAAAAAAAAAAAAAAAARkGYCgAAAAAAAAAAAAAAABgFYSoAAAAAAAAAAAAAAABgFISpAAAAAAAAAAAAAAAAgFEQpgIAAAAAAAAAAAAAAABGQZgKAAAAAAAAAAAAAAAAGAVhKgAAAAAAAAAAAAAAAGAUJhqmqqqrqurxqjpUVXuXeH5uVX1keP5IVW0f5rdU1Weq6ptV9d5F71xRVV8c3vn1qqpJfgMAAAAAAAAAAAAAAACwMUwsTFVVm5PcmeTHklya5Cer6tJFy25M8nR3X5LkV5PcMcw/l+SWJL+wxNbvS/K2JDuG66ozXz0AAAAAAAAAAAAAAACw0UzyZKpdSQ519xPd/e0k9yfZvWjN7iT3DOMHkry+qqq7n+3uz+ZEqOpvVdV3Jzmvuz/f3Z3k3iTXTPAbAAAAAAAAAAAAAAAAgA1ikmGqC5I8ueD+yDC35JruPp7kmSRbVtjzyAp7AgAAAAAAAAAAAAAAAPwdkwxTzVRV3VRVB6rqwLFjx2ZdDgAAAAAAAAAAAAAAADBjkwxTHU1y4YL7bcPckmuqai7J+UmeWmHPbSvsmSTp7ru6e2d379y6desLLB0AAAAAAAAAAAAAAADYaCYZptqfZEdVXVxV5yTZk2R+0Zr5JDcM42uTPNzdvdyG3f1nSb5eVa+tqkry1iQfP/OlAwAAAAAAAAAAAAAAABvN3KQ27u7jVXVzkoeSbE7ywe4+WFW3JTnQ3fNJ7k5yX1UdSvK1nAhcJUmq6nCS85KcU1XXJLmyux9L8vYkH0rykiSfGC4AAAAAAAAAAAAAAACAU5pYmCpJuntfkn2L5m5dMH4uyXXLvLt9mfkDSS4/c1UCAAAAAAAAAAAAAAAAY7Bp1gUAAAAAAAAAAAAAAAAATIMwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKAhTAQAAAAAAAAAAAAAAAKMgTAUAAAAAAAAAAAAAAACMgjAVAAAAAAAAAAAAAAAAMArCVAAAAAAAAAAAAAAAAMAoCFMBAAAAAAAAAAAAAAAAoyBMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAAAAAAAAAAAwCsJUAAAAAAAAAAAAAAAAwCgIUwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjIIwFQAAAAAAAAAAAAAAADAKwlQAAAAAAAAAAAAAAADAKAhTAQAAAAAAAAAAAAAAAKMgTAUAAAAAAAAAAAAAAACMgjAVAAAAAAAAAAAAAAAAMArCVAAAAAAAAAAAAAAAAMAoCFMBAAAAAAAAAAAAAAAAoyBMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAAAAAAAAAAAwCsJUAAAAAAAAAAAAAAAAwCgIUwEAAAAAAADrSlVdVVWPV9Whqtq7xPNzq+ojw/NHqmr7gme/OMw/XlVvXDB/uKq+WFWPVtWB6XwJAAAAAAAwbXOzLgAAAAAAAABgtapqc5I7k7whyZEk+6tqvrsfW7DsxiRPd/clVbUnyR1Jrq+qS5PsSXJZku9J8umq+r7ufn5470e7+6tT+xgAAAAAAGDqnEwFAAAAAAAArCe7khzq7ie6+9tJ7k+ye9Ga3UnuGcYPJHl9VdUwf393f6u7v5Lk0LAfAAAAAAAwEsJUAAAAAAAAwHpyQZInF9wfGeaWXNPdx5M8k2TLCu92kt+qqi9U1U0TqBsAAAAAADgLzM26AAAAAAAAAICzwA9399GqemWST1XVH3f37y5eNAStbkqSiy66aNo1AgAAAAAAp8nJVAAAAAAAAMB6cjTJhQvutw1zS66pqrkk5yd56lTvdvfJv3+R5GNJdi31z7v7ru7e2d07t27detofAwAAAAAATJcwFQAAAAAAALCe7E+yo6ourqpzkuxJMr9ozXySG4bxtUke7u4e5vdU1blVdXGSHUl+v6peWlUvS5KqemmSK5N8aQrfAgAAAAAATNncrAsAAAAAAAAAWK3uPl5VNyd5KMnmJB/s7oNVdVuSA909n+TuJPdV1aEkX8uJwFWGdR9N8liS40ne0d3PV9V3JflYVSUneqgf7u5PTv3jAAAAAACAiROmAgAAAAAAANaV7t6XZN+iuVsXjJ9Lct0y774nyXsWzT2R5DVnvlIAAAAAAOBss2nWBQAAAAAAAAAAAAAAAABMgzAVAAAAAAAAAAAAAAAAMArCVAAAAAAAAAAAAAAAAMAoCFMBAAAAAAAAAAAAAAAAoyBMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAAAAAAAAAAAwCsJUAAAAAAAAAAAAAAAAwCgIUwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjMJEw1RVdVVVPV5Vh6pq7xLPz62qjwzPH6mq7Que/eIw/3hVvXHB/OGq+mJVPVpVByZZPwAAAAAAAAAAAAAAALBxzE1q46ranOTOJG9IciTJ/qqa7+7HFiy7McnT3X1JVe1JckeS66vq0iR7klyW5HuSfLqqvq+7nx/e+9Hu/uqkagcAAAAAAAAAAAAAAAA2nkmeTLUryaHufqK7v53k/iS7F63ZneSeYfxAktdXVQ3z93f3t7r7K0kODfsBAAAAAAAAAAAAAAAArMkkw1QXJHlywf2RYW7JNd19PMkzSbas8G4n+a2q+kJV3TSBugEAAAAAAAAAAAAAAIANaG7WBazBD3f30ap6ZZJPVdUfd/fvLl40BK1uSpKLLrpo2jUCAAAAAAAAAAAAAAAAZ5lJnkx1NMmFC+63DXNLrqmquSTnJ3nqVO9298m/f5HkY0l2LfXPu/uu7t7Z3Tu3bt162h8DAAAAAAAAAAAAAAAArG+TDFPtT7Kjqi6uqnOS7Ekyv2jNfJIbhvG1SR7u7h7m91TVuVV1cZIdSX6/ql5aVS9Lkqp6aZIrk3xpgt8AAAAAAAAAAAAAAAAAbBBzk9q4u49X1c1JHkqyOckHu/tgVd2W5EB3zye5O8l9VXUoyddyInCVYd1HkzyW5HiSd3T3wAqgkgAAIABJREFU81X1XUk+VlUna/9wd39yUt8AAAAAAAAAAAAAAAAAbBwTC1MlSXfvS7Jv0dytC8bPJblumXffk+Q9i+aeSPKaM18pAAAAAAAAAAAAAAAAsNFtmnUBAAAAAAAAAAAAAAAAANMgTAUAAAAAAAAAAAAAAACMgjAVAAAAAAAAAAAAAAAAMArCVAAAAAAAAAAAAAAAAMAoCFMBAAAAAAAAAAAAAAAAoyBMBQAAAAAAAAAAAAAAAIyCMBUAAAAAAAAAAAAAAAAwCsJUAAAAAAAAAAAAAAAAwCgIUwEAAAAAAAAAAAAAAACjIEwFAAAAAAAAAAAAAAAAjIIwFQDA/2/vvsMmqeq0j983M+SopEVFBllAgjCSlCSgCLiiwEoYUGFWzKCAL7j6YhhWcUVZUUQMKAwiEiQISl6CBMnMDDAEyUqQpPI6iCjD7/3jnKare7qfp5/O4fu5rrnm6eqq6jqnTp2qEwsAAAAAAAAAAAAAAAAAAADASGAwFQAAAAAAAAAAAAAAAAAAAAAAAICRwGAqAAAAAAAAAAAAAAAAAAAAAAAAACOBwVQAAAAAAAAAAAAAAAAAAAAAAAAARgKDqQAAAAAAAAAAAAAAAAAAAAAAAACMBAZTAQAAAAAAAAAAAAAAAAAAAAAAABgJDKYCAAAAAAAAAAAAAAAAAAAAAAAAMBIYTAUAAAAAAAAAAAAAAAAAAAAAAABgJDCYCgAAAAAAAAAAAAAAAAAAAAAAAMBIYDAVAAAAAAAAAAAAAAAAAAAAAAAAgJHAYCoAAAAAAAAAAAAAAAAAAAAAAAAAI4HBVAAAAAAAAAAAAAAAAAAAAAAAAABGAoOpAAAAAAAAAAAAAAAAAAAAAAAAAIwEBlMBAAAAAAAAAAAAAAAAAAAAAAAAGAkMpgIAAAAAAAAAAAAAAAAAAAAAAAAwEhhMBQAAAAAAAAAAAAAAAAAAAAAAAGAkMJgKAAAAAAAAAAAAAAAAAAAAAAAAwEhgMBUAAAAAAAAAAAAAAAAAAAAAAACAkcBgKgAAAAAAAAAAAAAAAAAAAAAAAAAjgcFUAAAAAAAAAAAAAAAAAAAAAAAAAEYCg6kAAAAAAAAAAAAAAAAAAAAAAAAAjAQGUwEAAAAAAAAAAAAAAAAAAAAAAAAYCQymAgAAAAAAAAAAAAAAAAAAAAAAADASGEwFAAAAAAAAAAAAAAAAAAAAAAAAYCQwmAoAAAAAAAAAAAAAAAAAAAAAAADASGAwFQAAAAAAAAAAAAAAAAAAAAAAAICRwGAqAAAAAAAAAAAAAAAAAAAAAAAAACOBwVQAAAAAAAAAAAAAAAAAAAAAAAAARgKDqQAAAAAAAAAAAAAAAAAAAAAAAACMBAZTAQAAAAAAAAAAAAAAAAAAAAAAABgJDKYCAAAAAAAAAAAAAAAAAAAAAAAAMBI6OpjK9k6277V9v+3P1fh+Udtn5O9vtD2l8N3n8/J7be/Y6D4BAAAAAAAAAAAADDfaIQEAAAAAAAAAQLM6NpjK9iRJ35P0LknrStrb9rpVq+0v6c8R8a+SjpF0VN52XUnTJK0naSdJx9ue1OA+AQAAAAAAAAAAAAwp2iEBAAAAAAAAAEArOvlmqs0k3R8RD0bEPySdLmmXqnV2kXRy/vssSe+w7bz89Ih4MSIeknR/3l8j+wQAAAAAAAAAAAAwvGiHBAAAAAAAAAAATevkYKrXSvpD4fOjeVnNdSLiJUnPSVp+jG0b2ScAAAAAAAAAAACA4UU7JAAAAAAAAAAAaJojojM7tneXtFNEfDh//qCkt0TEgYV17szrPJo/PyDpLZJmSLohIn6Wl/9E0kV5szH3Wdj3RyV9NH9cW9K9bQ8kVpD0TK8Pok8QF5WIjzLiohLxUUZcVCI+yoiLSsRHGXFRifgoIy4qER9lxEUl4qOMuKhEfJQRF5WIjzLiooy46JzVImLFXh8EgMFAO+TQGtb7LOEaLMMaLml4w0a4Bsuwhksa3rARrsEyrOGShjdshGuwDGu4pOENG+EaLMMaLml4w0a4Bsuwhksa3rANa7h6oWY75OQO/uBjklYtfH5dXlZrnUdtT5a0rKRnx9l2vH1KkiLiR5J+1OzBY3y2b4mITXp9HP2AuKhEfJQRF5WIjzLiohLxUUZcVCI+yoiLSsRHGXFRifgoIy4qER9lxEUl4qOMuKhEfJQRF2XEBQD0Ddohh9Cw3mcJ12AZ1nBJwxs2wjVYhjVc0vCGjXANlmENlzS8YSNcg2VYwyUNb9gI12AZ1nBJwxs2wjVYhjVc0vCGbVjD1U8W6uC+b5a0pu3VbS8iaZqk86vWOV/Sfvnv3SVdEelVWedLmmZ7UdurS1pT0k0N7hMAAAAAAAAAAADA8KIdEgAAAAAAAAAANK1jb6aKiJdsHyjpEkmTJJ0YEXNt/5ekWyLifEk/kXSK7fsl/UmpUUJ5vTMl3SXpJUkHRMR8Saq1z06FAQAAAAAAAAAAAEB/oR0SAAAAAAAAAAC0omODqSQpIi6UdGHVsi8V/v67pD3qbHukpCMb2Sd65ke9PoA+QlxUIj7KiItKxEcZcVGJ+CgjLioRH2XERSXio4y4qER8lBEXlYiPMuKiEvFRRlxUIj7KiIsy4gIA+gTtkENpWO+zhGuwDGu4pOENG+EaLMMaLml4w0a4Bsuwhksa3rARrsEyrOGShjdshGuwDGu4pOENG+EaLMMaLml4wzas4eobjoheHwMAAAAAAAAAAAAAAAAAAAAAAAAAdNxCvT4AAAAAAAAAAAAAAAAAAAAAAAAAAOgGBlMNGdvL256d//3R9mOFz4t06Rhm2D60zndh+2eFz5NtP2371/nz9Px5tu27bH+kjcc1vxAXs21Pade+J3AMU2y/YHuW7btt32R7euH7joU/77/tcWB7pu2H8v5us715g8dwp+1f2F6iavlc23Ns/x/bCxW228z21bbvzfH349K27TReGp3Afq6yvUn++0Lby9VYp+61Umsd24vZvsz2jIkcS7vZvtL2jlXLDs7p4HNt/J1X4rBf2F7Z9s9tP2j7VtvX295tnG1qnv8mf7+YR9xj+5B27LeJ45hX47iOa3GfD9teoYnttrF9fdWyybaftP2aVo5pjN9sa15q+3uFfP+Fwn53r1pvx8J383J+ONv2T21vm/Ov9xTW/3Vefm5e737bzxX2sUUrx91kWGveAya4j+Vsf7ITx9cpvbz/TvQ+3S014mTC949G7qP9yPbh+Xnn9hz2t+T7aM3rodk8Nv9OKX6L8R318pxup5deXhs1jqHes2nTabTB4205X+zAMW1buF/cbfvLYy1v82/3Q5ltIdvH5nNyh+2bba9u+yDb3y6s90Pb/1v4/Cnbx07wt4al/D7L9n22L3Hh+aLTeUq38tO87dpO5ZNS+v9RXr6E7VNzWrnT9rW2l2olXE0eXyfy07favrEQ5hkNbLPAOWn1OBrR7vDb3s/2aVXLVsjpfdEa60+x/agLdRh5eSldzihc33fafm8rx9eoTuWphbzg62Oss8D1lo/h9HYcw0R5nDqnDv/2FNv7FD7Xipu+q/8AAGAsfVJ263h72zCWOTwCbWAdKh91tC7Wg18/sLJTe8ScfL1d2MvzUFi/Xr3fv9g+3fYDTm19F9resHCsfyr8zv/anurUHljKD/Zq4FjrlkFsb+IJ1iFNlKva8VrYT6fP42zbv636/pe2b6haVixXl/4tV/j+2/n7lst57conG/ytqfn3dmrzfkeuDOwB7b/Q7mvMfVqn1aG8pGt1lx04T432DfiPwnf/cHoenG3766581i39Wzef07D9qcJ+jnPhObkbYcz7rJnfu85z+jhheqGw/g8mmqd5iPq3dOJcTZRbbEfrkzA0Up6se/+qdf4bTVceoH5nHvxySimsE2rH7LM0WpH3jZd2q/YxLOev6Xqdqt8byP5qnUiP9dJ+A8fQ1j417qP7c2H7fV2+v80aI/3283mZUJ1HYX/LOJUNWjoHg2pyrw8A7RURz0qaKqWbkaR5EXF0Tw+q0vOS1re9eES8IOmdkh6rWueMiDjQ9kqS5to+PyKebMNvvxARU2t9YduSHBEvt+F3xvNARLw5/+4bJJ1j2xFxUv6+U+GXOhcHh0XEWbZ3kPRDSRs0cgy2T5X0cUnfqlq+kqSfS1pG0pdtryzpF5KmRcT1eZ3dJS0t6W9NHO9YGkmjExIR/9bqQeWH2LMl3RoRM1rdX4tOkzRN0iWFZdMk7RcRV1evbHtyRLzUrYPrlHyN/FLSyRGxT162mqQxKw1rnf8Wr7dSHrG8pHttnxURf2hiPz3T5jRxjaTX2V4tIh7Jy7aXNDciHm/Tb1Rra14aEQfkbadI+nW9fUfEJcrXne2rJB0aEbfkz9tKelTS4ZJ+VbXdboV1Do2InRs9tg6odw+YiOUkfVLS8W0+tk7qyf3X9qRG1uuRunEyzHJhd2dJG0XEi7kQv4ikMyT9TG18romIIyUdmX93XnV818pzbO+s7qaXgXg27bAx88UePkddExE7215S0mzbv6q3PCJua+Pv9kOZbS9Jr5G0QUS8bPt1SmWE6yS9v7DehpIm2Z4UEfMlbSHpvIn80DCV3yXJ9nZK5dvtIuLu/H1H8pRu5qfZsZKOiYjz8u+/KS8/SNKTEfGmvHxtSf9s8283ohPXzsmS9oyIOfmZYu2xVh7jnHRDu8N/rqT/sb1ERJTS0u6SfhURL1avHBEP2/69pK0l/Sb/7hslLR0RN9p+l1L6Odr2OpKusb1SF/KzTuWp75T0O0l72P58RMR4G+RwT5K0te0lI+L5Jn63FXXrnLrw21Mk7ZN/EwCAYdEPZTepg+1tQ1zmGIU2sIGrix2C+oH/knRZRHxHkmxvIOlt/VjvZ/sYpTLfyRExLX+3oaRlCuvOVKo3PSt/XkvSvhFxn1Mn51ttXxIRf2nwtyvKILld55Ymwl6hS/WGHb2eauxzOUkbS5pn+w0R8WDh62NqXRdOHep3k/QHSdtIurKJ4ylqez45hr0lXZv/v7iN+x3FMvCg9l8YlTqtQa+77FXfgJMknZTXfVjSdhHxTP48XYW6+cLxTJH0lKSDbP8wIv7R4GF1Nb9Xjef04vKq35+i9Nw/1fZkSVdI2lXSOU0cT1PanGe02r+lH8pirbajfagPwiCNX55suxy+gel3NgTlFKm5dsx+uM6k2nnfbWow7Q7T+WumXqeGQe2vRp+aOtr9TJ+ffQ+WtENEPO40GcG+dVbvy/PSTJ1HwVckLVB2GhW8mWoE2P6I0ywAc2yf7fIIxJm2v2/7BqfR7tvaPtFp1PLMwvbft32L0wwdRxSWP2z7CKdRkHfkQnXJuk4j8h+0/emqQ7pQ0rvz33srVWwsICKekvSApNVaj4UFOY3Uvtf2TyXdKWlV24fluLq9KqyH2/6d04xvp7k8S9pVLs86sEIuwMr2JNvfLOzrY3XC+KCkz0iqjqOOhz8fZ8txUOVqSf9qeynblxfSxi51DuEaSf9avTCH/aOSDsw3lwOUMvjrC+uc1cLD0XjqplHbS+br5Can0ce75OWLO43ovdv2uZIWL2zzyojnYjyqUKFT7zrNJis10N0XER15K8IEnSXp3c6zFDhVXrxG0hrOI5Nz/vID2zdK+obtNXJec4ftrzqPLM/5ziszebmBWXF66O2S/hERPygtiIhHIuK7rhoZ7/xGoPz3wzl/qHW9jZW/ltLMJk4DZyrkQs/9klbpUHibYnvFnIZvzv+2zMtn2D7F9nWSTnGaAePSHPYfS3JhH790Ghk/1/ZHC8vn2T4yXyc32F45P3ieqVQhXjJNde4tndCBvLRuXjOOOZKes/3O9oSs465Rume8Op/z2/N53UB6Jc2c6AWfJ76ulN/Mzvda5/9LM0OMOytkr3UgzVytfD/N+cdRtm+TtEe99fqR7WVzvKydP5/m8mxk++a4mWP7lBrbtvS820WrSHqm1GiVGz12V7qPXmn7ynzc/5HP+02StixtXCuPdZqB6GFXzsR5n9Ng9FZ0Pb106tpwi8+mXVTKF7e1fY1T49FdTjMzn+Ty7DfbSamxzPY5ti/O5/wbpR3Z3tvlWauPKiyfl/PMuU4z625WyGcXqKiP1Nn9VlXFS73l7daONOGJl9lWkfREqXIrIh6NiD9Lmi1pLafn/mUlvZCXlTq4baHUUNRqmAey/C5JEXGlpB8pleWqtTtP6XZ+uorSwPVSWO8oHMdjheX31uqY0G1tyk9XkvSEJEXE/Ii4K68/o5jn5nxmimqckwYbf9uu1fBHxP9T6kDynsJup0k6rVbayd+XOu0U11/gLUy5ge4lSU3PitasNqULKeUF35H0e0mbF7apeb0VtjlF0qWSGinbdEx1nZPHLs/vXwqT7RNcrvN4j9Psx7Oc7qcr5+XbuDyz2yzbSyuVX7bOy3rydmkAADqtHc8Z7r/2tmEuc4xUG1gbn4NLulIXWy/O3J/1A9Xp9/Ya4emXer/tJP2zqq1vTkRcU2c7RcTvIuK+/PfjSp3UV6y3fo3tq8sg2zqVO+rmCbXykPx9dVvbevl6nZ3jcc1Gj6sZHTiP1f5dacLA01VZxh7LtpLmSvq+Utpsh7Hyyc2c3iAxy/ZvXW7PmOJUp3tb/rdFXr6K7atdnjF867zcSvnGdEnvtL1Ym469wgiVgYei/0Kr19ig1Gm1KS/pWd1lJ/JC12kvb9LTki6XtF+zO+hCfi9p4s/pucP0b9XGtoZ614YHpH9LO86Ve9yO1qMwVBirPNlmA9/vzINVTqkOb1PtmL1Oo2PlfRNNu4N6/jrQj3pg+6u1Iz1W6cs+Ne7y/Tl/9XmlSeofl6SIeDEiTmjwePvlvEy4ziMf08aSVlZqxx1JDKYaDedExKYRsaGkuyXtX/juVUodHw6RdL6kYyStJ+lNtksjJw+PiE2URjlu48pC2zMRsZFS5VTxon6jpB0lbab0ZqGFC9+dLmmaU4XQBpJurHXQTiOn36D00NYOi7tcmXNuXrampOMjYj2lSv018zFPlbSx7bfljGJaXvZvkjZt4Lf2l/RcRGya1/+I7dXrrHubUnxV6ED4pc7HwXsk3SHp75J2y2ljO6XZb1xc0WnU/Lvy+gvID3uTlCpB1lfqrNktY6XRwyVdERGbKYXtm04z839C0t8iYh2lWZ02rt7pOPE41nX6WaXC1MHtCmArIuJPkm5SOn9SCtOZkqpno36dpC0i4jNKHa2+E2n2xEc1mNZTul5b8cr1FmmWmbHy1zHZfr2kxSQt0CjVBcW8ZLbSrIMl31GaoWpTSe+T9OPCd+tK2j4i9la6Tq7Nec+5kl5fWO9DEbGxpE0kfdppNhRJWlLSDfk6uVpS6RW+r1TyOs0K8G9Ks1h2SjfuJ/XymvEcKekLTYWqi6ruAUdImhURG0j6v5J+Wli11vPE55RnYImIw5Qa16YqzWy0vVJc9dUgQ3Xv/lvybERsFBHVjR3V6/VSRT5ie6+IeE7SgZJm2p4m6VURcYLt9ZTS9dvz9X9Qjf21+rzbLZcqFZp/Z/t429tExLGSHleaUW67nH6PUOqAs5VS3lmyQB6bK6rPU5p9U7bfIumRNgw670Z66edn0wXSaLsCXUuN399I0kERsZbSxAKRn6P2lnSyy43rU5VmgHuTpL1sr+o0Q+5RShXyUyVtanvXvP6SSveX9ST9VdJXlWZ52k2V9/PScS0v6a1KHRLGXd4G/VBmO1PSe/Ix/I/tN0uvVFTPyuu+VamMcIOkLWy/VpKjPW8LHcjye0HN8q3an6d0Oz89RtIVti+yfYjLHZ1OlPSfTp1ovuoOd1oaQyeunWOUZiQ81/bHPH6nngXOSRvDN55OhL9YxniNpLWUZv6rV945U9KuOT+XUt68QKNPTlcvK3Vq6LS2x0tOB9srdW47Tbmj2jjXm5Ti4/TiNr1UVedUUz7vX1TK87dUZd52raS3Rpr98XSlehsp5a0HRJrdbWulDgOfU3qz49SIOKbdYQEAoEf6oexWSzvb24a5zDHsbWDDUhc7SPUD35P0E9tXOnUKeo36t96vpTZn25spvcnkgYlsV6sMMk6e0Ghb28eV2l+nKrWptbsNtlPn8ZuF/Z5aWF7qQFmr7HhIYZsra2xzrtJgmoXVurHS+z2Sts7lwS9J+lpe/pSkd+a0uJfSGwel9JamS/I52lCpU7mUOpQ/FBEPSLpK5U6kbTcKZeAB7r8wKnVag1532Y1n37Hay+vZy5XtWIsXvjtK0qEuvzmzV2Gsl99LqvmcPlaY5NTp/x2a+DPWMPVv6YeyWKvtaP0Qhlqqy5N7VaWbTRr4vfEMQ7+zQSqn1NJIO2ZfpdEG8r56YaplIM9fjftF0zx4/dX6tWxdfWzN9Knpt/vzROoL+vW8TLjOw+ltz/+jyut25EwefxUMgfVtf1XpNYNLqfLV1r+KiLB9h6QnI890Znuu0iu4Z0va02lk5mSlkejrqvwQVXpt7q1KN4eSCyLN6PGi7aeURi0+KqXZqJxm/thbaXRxtb1sbyXpRUkfyxUf7VDxar18DI9ExA150Q7536z8eSmlDG1pSedGfg22y68XHssOkjawvXv+vGze1+9qrOuqz50Kv9S5OPim7S8oVZzsrxSmr9l+m1KFymuV0sAflW8kebtrJP2kjeFri3HS6A6S3uvyiODFlG7Cb1OuFM3b1ypobK368TjWdXqtUuFyrYiolYZ6oVS4Py//v7/Ks4mU/CLSq5ql9LBd6sT7c0n99NrYptj+nlID6T+UGooaUbzepLHz13r2ytfWGyUdGBF/n9iRt0V1XjJd5YqD7ZVmlSh9vYztpfLf50d6Na+Urpl/l6SIuMD2nwv7/7Tt3fLfqyrlQ88qxXVpJrBblTp/KyJucRqJv7akdSTd2Oa8s1o37if18pq7628iRcTVtpXvI/2o1j3gRqWCjyLiCqdZI5bJ69R6nqi2laTTcn7zpO3fKBU2Grlfd0u37r8lZzS4Xi/VfOVxRFxmew+lfHXDvPjtSveUZ/I6ta7vVp93uyIi5uXC8dZKBd0zbFfPuPsWSVdFxNP5OM9QauyS6uexZyg13J6kdF+uTgMT0c300s/Ppt16JXmt399C0k0R8VBevpWk70pSRNxj+xGV08TlkQYiyvZdSrMxLa/KNHSq0n33l0r30ovztndIejEi/pmvjymF49ra9iyl+Pp6RMx1mv1sgeXtiwpJfVBmi4hL8zPF2/O/y23vERGXK836tYXSDNzXS7pPqWL16fxdOwxa+b1adfm2I3lKt/PTiDjJ9iWSdlJ6q87HbG8YEbNzBf4OeZ8329480kyt3dT2ayci/ivnHzsodfzZW2m255rqnZOImNmG8I2nE3nHBZKOz8+le0o6OyLm266ZdiLiSdt3SnqH7SclvRQRdxb2d4jtDygNZN0rIqo783RCJ+JlZ0lXRsQLts+W9EXbB2uM681ptsVnIuL3th+TdKLtV3e4zNYOm0n6Tek4bf9C5TzkdUppfBWljpSle/Z1kr6Vr51zIuLRqjYNacGOXOMtBwCgH/W87KYOt7cNc5ljBNrAhqUudmDqByLikpxOd1LqwDNL/Vvv93E1KT//nyJpv8hvYmiDevW6jba1XS/pcNuvUyqD3Nem4yrp1Hk8LCLOKi5wmg18TaWOcGH7n7bXL5Stj4mIo6u2WUSpA9pnIuKvTm8c2lHltsSmjJNPLqs02dWaSuW4UmfQhSUd59TRdL7K94OblcrBC0v6ZUSU0uTeKr/953RJ+6qzE0WOZxjKwIPYf2FU6rQGve6yG8++W6lGe3mkt43Vc0ZEHFhcUDq/EfFgzhP3aSB8Uhfz+2yB5/R87PXCtEa+p4ek8yLiogbDVS980zW4/Vt6XhZrQztaz8OgxsqTFenRNd7s1CoPZr+zgSmn1NFIO2a/pNEF8r58LOOFaSyDdv7a2Y96UPur9WvZeoFja0PYpquH9+cWj32K+uO8NFPn8UlJF9Ypz40MBlONhpmSdo2IOTnD2bbw3Yv5/5cLf5c+T3Ya6XyopE0j4s9Or21crMb281WZnor7qv5OSjeMo/OxLF/13QKFow56vvC3Jf13RPywuIJTp5B6XlL5DW/FeLGkT0VE8YGjlGlWe7MqO8d3M/xS63EgVRWAczpbUdLGkTpnPqxy/DR0E82V7vOVZnGaqzTL3XnjbddG9dKoJb0vIu4trtyGG8lM1b9Or5Z0sqSLbG8VEU+0+mNtcJ6kY2xvJGmJiLjVdnVl5PM1tqtWvIakyuuo38xVfoCWpIg4wOmVyLeo8XC8Eifj5K/18hYp5xFOHdIutX1+RPyxyTB1wkJKs5NVFLbzNTJumnDqqL29pM0j4m+5QqIUB/8sVNBW31tKFeTrqMlXoLeoHXlpxeqqkdc0qPR2qpea2LbTFrgHjJN/jvc8Mcjafv+ts++x1us7TjNerCPpb0oz3zQ6G+BMNfm828LhNiVXpFwl6apcEbXfBDavl8der/SK5xWVGv++2sIh9jq99OWzaQfVyxcbeY6SJp5PFu+lr1wTEfGyyzNPSmn20J1rbF9veSd1tcwmpVe2S7pI6fn7SaXr6nKlDgIfz/v5nlIl1rpq72CqmRqs8nu16vJtx/KUbuenEfG40qzwJ+YOButLujUi5ik1EJxj+2WlTjzdHkxVS8v5aaTZkb9v+wRJTzvN2FW37FPnnMxsIQytaCn8kQYLXaw0Q/g0SZ/JX9VMO1mpTPKkFiyTLNDpq0daTRd7S9oq30ullCe8fZzf3FvSGwvbLKNUtj6hwWNuu6o6p2bqJb4r6VsRcX4uw86QpIj4uu0LlPKB62zvWGPbZ5Wec4teLemZiYQBAIA+NHTtbUNe5hi1NrBBrIudqQGqH8id2n4u6ee2f600c3q1ntf75Y6Au2uCcoe6C5RmYr9hvPVrbF8sg6xT+KpentBQW1tE/Nyps/y7JV1o+2MRccVEj2+C2t0eVrKIKrZJAAAOWklEQVSnUlnpoRzOZZTKk4ePsc2OSh1B78jbLKH0dqSWBlNl9dL7V5QmGdkt34uuyssPUaoP2FDp/P1demUSxLcpnaOZtr8l6VSlcvEutg9XisflbS8dEX9tw7FXGKEy8LD0XxiVOq1Br7vsVF7YTl+TdJak3zS5fSfDONHn9Ac62JY3DP1bhqEdrR/Lk50wDP3OZmqAyik1NNuO2Ys02mjeN5G0O1ODdf7a2Y96mPqr9bxs3QW9uD+X+qg3W6bu+Xlpss5jc6VJlT+pNOhrEdvzIqJ6UqmhttD4q2AILC3pCafZbt4/wW2XUbrIn3OaDehd46zfqBMlHRF5BHOfuETSh5xHsNp+re2VlCrxd7W9uO2llV6hV/KwUgYqVWZCl0j6RI5z2V7L9pLVP5gffI5Wnmm+DzQTB7UsK+mpnIFvpzRDfsNyhfEPJB2Xb1zHSdrP6dXhpXX+PafJTqmXRi+R9CnnO7Pz64qV4mifvGx9pVePVhsrHse8TiPibKW0crHt5ZoPVnvkxrsrleKpkYL9DSoXCKcVlj+iNIp80RyuWo0r/eIKSYvZ/kRh2RL5/4clTbW9kO1VlWbvGs9Y+evDKuct71MNEXGL0gx4BzUagC65VNKnSh9cftVvteI18y6VK9KXlfTn/CD7RqXXfzfiNEkfUOrI182Bl7W0Iy+tl9eMKyIuVYrPhl/f3WPXKOd7uTDzTIw9y9ZflfLM4vZ72Z6U7x9vk3RTh461U9p1/x0GhyhV9Owj6aR8X7xC0h65EUS2X11ju1aed7vG9tpOM1aWTFW6FxbT9Y1Kr0RfPodnj8L6NfPY/Lx0rqRvSbo7Ip7tXCi6qi+eTftAMZ9cS2lG6LEG296klIZWsD1JqdNBs41l/abjZTbbG9l+TV62kNL99JG8zfVKzyYrRsRT+dp7WmnW8OvaFMaBLb/b3kbSR9WFARLdzk9t71RIK/+iVJn/mO0tbb8qL19EqVHwEfWfCV87tt9dehZVmilrvqS/KF1TG+V1NpK0ev673jnpB83eT05T6nCystL1L41d3jlHqQPRXirPLt3PJhQvTp0Ht5b0+oiYEhFTJB2gdJ+peb3lfHRPSW8qbLNL3qYnatQ5Paza5fmblcL0KqcBx8Wy+bKSHst/v9Kp2vYaEXFHRByVt3+jFiy/3Cxpy5yXyKnBelFJf2hvSAEA6KmBb28bgTLHKLeBDUpd7MDUD9h+u+0l8t9LS1pD4791plf1fldIWtRpNvXS8W9ge+t6G+Rr71xJP21moFyNMsgrxqjXbaitzWmQzIMRcaxS21i324XaeT3tLWmnQtlxY1W2Kdfb5sOFbVaX9M5SemxRvXyyWB6cXrX8iUhvLfugpEmSZHs1pZn9T5D0Y6U6lXdIuj0iVs3HvprSW6l2U5uNUhl4SPsvjEqd1qDXXbYjL5xoe/m4IuIeSXeN87uNGpTnp2YMW/+WYWhHG/jy5BiGod/ZwJRTqrl97ZhdT6P1NJF2B/b8dcgw9Fcb1j41vbg//7fSm6FKZZdFbH+4yeMfmDqPiHh/RLw+l6kPVar7GKmBVFJ/jZZE53xRqUL/6fz/0mOvXpZHIc+SdI9ShUZbOohFxKOSjm3Hvtol0mto15F0fS5zz5P0gYi4zfYZkuYozdZzc2GzoyWdmTOfCwrLf6z0esvbcgH+aZVfEb5GjtPFlG6wx0Zjr47uuCbjoJZTJf3KaTaXW5TSz3hKrx1cWGk0/ilKFceK9GrxaZKOzjeVl5VuLBdPNIyNGiONfkXStyXdnguFD0naWdL3lTp9363UCfzWGvscKx7HvU4j4vv5YfR82ztUj7zugdOUKvjHq8iWpIMl/cxpdq2LJT0nSRHxB9tnSrpTKS5n1d9Fb0VE2N5VaUarzyqdq+cl/adS3viQUmXU3ZJua2B/Y+WvR0j6ie2vqDyjWS1HKeUzX4sOzFTWpE9L+p7t25WeM65W7VeIHiHpNKcR8b+V9Pu8/GJJH8/X0r1KFdnjioi7bT+vNFNno2/z6Ig25aX18ppGHaneDypr1Ayl2VZvV3ob0ZgzvEbEs7avc5qd9SJJn1WaJWGOUgPtZycwa05faOP9d5AUXzcspWv/JEkflrRZRPzV9tWSvhARX7Z9pKTf2J6vdK+YXrW/pp93u2wpSd91aoB7SdL9ShV2eyt1Fnk8IrazPUOpsvkvSq9LLxkrjz1DKY1M70I4uqJPnk1LLu5hhcHxSrMr3qGUbqZHxIuuM1NSRDxh+3NKDcdWegX9oNwTxtSlMttKkk6wvWhe7yalyR2UZ9l6WmlWopLrJW2Zf7sdBq38vpftrZQaex5SmsG8G29l6nZ+uoOk79gulcEOi4g/2t5B6fq00oRFFyh1fukrTV47H1Qq+/xNKY7fHxHzbZ8tad/8HH+jpN/l9eudk55r4X5ymaSfSvpJocNb3bQTEX9xetPAv0TEg50OV6uaiJfdJF0RadbRkvMkfUPSJ5Se66uvt60lPRbpLQslVyt1Sloluvfm7bp1TqpTno+Ix2x/Tek+8CelvPW5vM0MSb+w/WelhonV8/KDc2PGy0r3iovy3/Ntz5E0MyKOsX2Q0qzxCynF+965wx0AAENhSNrbhrrMMcptYANUFztI9QMbSzrOdmm29R8rtQGNta+e1Pvltr7dJH3b9n8qvT3oYaV2zHr2VOoUt7zTrNBSqp+bXX+TMcsg1WrlCY22te0p6YO2/ynpj0pvIZGkJWw/WljvWxFR7/eb1sJ5/KbtL1SFYzUV2gQj4iHbz7k82eohtj9Q2GYfSTupEC8R8bzta5U6pJ3RYtjqpfdvSDo5H3/xfnW8pLNt76vU3lFqq9xW0mH5HM2TtK+kLym1sRedrVS2/mkrx52Nchl4qPovjEqd1qDXXbbpnjZDE2gvz0p18yWflPR41TpHqg1puI35/XiDQxoJU7sNVf+WYWhHG6Dy5O1Obw6WpDMl3S5pulO/spK35meKUtiGod/ZIJVTpA60Y3YxjdbTStodtPPXaTM04P3VhrhPTdfvzxFxYa6X+t98PYbSYL8JG7A6D0hyVE5+A2AcuWFkXnTmVdYDgTjARDnNAPZCvmFPU6ok3aXXxwX0EnkpJoo0A9TGtYFqpAmgOaN+7Yx6+OsZ5XixvVREzHOalftcSSdGRHWHNwAA0KRRfs5A/yN99gfOw3DgPA6GUSsDD1P/hVG5xgY9nIN+/I0YhTAOi2E4V8MQBgw30ij6CemxP3Fe+gdvpgIAdENpVjorzdD4oR4fDwAAAAAAwFhm2N5eaYbHSyX9ssfHAwAAAABAp4xaGZj+CwAAAAAA3kwFAAAAAAAAAAAAAAAAAAAAAAAAYDQs1OsDAAAAAAAAAAAAAAAAAAAAAAAAAIBuYDAVAAAAAAAAAAAAAAAAAAAAAAAAgJHAYCoAAAAAAAAAAAAAAAAAAAAAAAAAI4HBVAAAAAAaZnu+7dmFf1Oa2Meuttdt/9EBAAAAAAAAAAAAGES0QwIAAADopsm9PgAAAAAAA+WFiJja4j52lfRrSXc1uoHtyRHxUou/CwAAAAAAAAAAAKA/0Q4JAAAAoGt4MxUAAACAltje2PZvbN9q+xLbq+TlH7F9s+05ts+2vYTtLSS9V9I384xya9i+yvYmeZsVbD+c/55u+3zbV0i63PaStk+0fZPtWbZ3qXEs2+b9nWX7Htun2nb+7kv5eO60/aPC8qtsH2P7Ftt3297U9jm277P91cK+P5B/e7btH9qe1Om4BQAAAAAAAAAAAEYV7ZAAAAAAOoXBVAAAAAAmYvFciT/b9rm2F5b0XUm7R8TGkk6UdGRe95yI2DQiNpR0t6T9I+K3ks6XdFhETI2IB8b5vY3yvreRdLikKyJiM0nbKTWELFljmzdLOljSupLeIGnLvPy4fDzrS1pc0s6Fbf4REZtI+oGk8yQdIGl9SdNtL297HUl7Sdoyz4g3X9L7G4oxAAAAAAAAAAAAAOOhHZJ2SAAAAKBrJvf6AAAAAAAMlBdyJb4kyfb6SpX9l+UJ1iZJeiJ/vX6eUW05SUtJuqSJ37ssIv6U/95B0nttH5o/Lybp9UoNJEU3RcSj+fhmS5oi6VpJ29n+rKQlJL1a0lxJv8rbnJ//v0PS3Ih4Im//oKRVJW0laWNJN+dwLi7pqSbCAwAAAAAAAAAAAGBBtEPSDgkAAAB0DYOpAAAAALTCSpX+m9f4bqakXSNiju3pkrats4+XVH5r7mJV3z1f9Vvvi4h7xzmmFwt/z5c02fZiko6XtElE/MH2jKrfKm3zctX2LyuVmyzp5Ij4/Di/DQAAAAAAAAAAAKB1tEMCAAAA6JiFxl8FAAAAAOq6V9KKtjeXJNsL214vf7e0pCdsLyzp/YVt/pq/K3lYabY1Sdp9jN+6RNKnnKdks/3mCRxnqcHiGdtLjfM7tVwuaXfbK+XffrXt1Sa4DwAAAAAAAAAAAACNoR0SAAAAQMcwmAoAAABA0yLiH0oNAkfZniNptqQt8tdflHSjpOsk3VPY7HRJh9meZXsNSUdL+oTtWZJWGOPnviJpYUm3256bPzd6nH+RdIKkO5UaQ25udNu8/V2SviDpUtu3S7pM0ioT2QcAAAAAAAAAAACAxtAOCQAAAKCTHBG9PgYAAAAAAAAAAAAAAAAAAAAAAAAA6DjeTAUAAAAAAAAAAAAAAAAAAAAAAABgJDCYCgAAAAAAAAAAAAAAAAAAAAAAAMBIYDAVAAAAAAAAAAAAAAAAAAAAAAAAgJHAYCoAAAAAAAAAAAAAAAAAAAAAAAAAI4HBVAAAAAAAAAAAAAAAAAAAAAAAAABGAoOpAAAAAAAAAAAAAAAAAAAAAAAAAIwEBlMBAAAAAAAAAAAAAAAAAAAAAAAAGAkMpgIAAAAAAAAAAAAAAAAAAAAAAAAwEv4/MbdAZFJSMa8AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 4320x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IGMK2P869VOG",
        "colab_type": "code",
        "outputId": "94741a0b-384b-46e7-8608-6fd95e9f386b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 289
        }
      },
      "source": [
        "print(news_features.index)\n",
        "print(std_features.index)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Index(['TamanhoMedioParagrafos', 'FrequenciaDeParagrafos',\n",
            "       'FreqPalavrasDeConteudo', 'MedidaK', 'Virgulas', 'GuiraudR', 'HerdanV',\n",
            "       'FreqTermosNaoTageados', 'Ponto', 'FreqPronomes', 'Exclamacoes',\n",
            "       'StdevTamParagrafos', 'FreqPalavrasErradas', 'PronomesPorPreposicao',\n",
            "       'FrequenciaStopWords', 'TamanhoMedioDasPalavras', 'StdevSentencas',\n",
            "       'FreqSubs', 'FreqVerbosPtcp', 'FreqAdv', 'DugastU', 'FreqPrep'],\n",
            "      dtype='object')\n",
            "Index(['FrequenciaDeParagrafos', 'StdevSentencas', 'MedidaK', 'FreqPronomes',\n",
            "       'TamanhoMedioSentencas', 'FreqPalavrasComunsTam2', 'DiversidadeLexica',\n",
            "       'FreqEAbstracao', 'MaasA', 'DugastU', 'Virgulas', 'FreqVerbosPtcp',\n",
            "       'FreqSubs', 'FreqTotalEntidadesNomeadas', 'FreqEPessoa', 'HerdanV',\n",
            "       'FrequenciaStopWords', 'FrequenciaDeHapaxLegomenaLocal', 'GuiraudR',\n",
            "       'TamanhoMedioDasPalavras', 'FrequenciaPalavrasDuplicadas',\n",
            "       'TamanhoMedioParagrafos', 'Ponto', 'FreqPalavrasErradas', 'HerdanC',\n",
            "       'FreqPalavrasComunsTam4'],\n",
            "      dtype='object')\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1C2UZePwRo72",
        "colab_type": "code",
        "outputId": "c01936d4-edff-4d1e-eeb8-201ab7a827b5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "# TODO: Finish the pipeline with feature selection + normalization + classification + hypertuning + metrics report\n",
        "X, X_test, Y, y_test = train_test_split(df_news.values, encoded_y_news, random_state=7, test_size=0.3)\n",
        "\n",
        "treeclf = ExtraTreesClassifier(n_estimators=400, random_state=11)\n",
        "\n",
        "clf = Pipeline([\n",
        "  ('feature_selection', RFECV(estimator=treeclf, step=1, cv=5, scoring='accuracy')),\n",
        "  ('classification', ExtraTreesClassifier(random_state=11))\n",
        "])\n",
        "clf.fit(X, Y)\n",
        "\n",
        "y_pred = clf.predict(X_test)\n",
        "\n",
        "score = accuracy_score(y_test, y_pred)\n",
        "print(\"Test score: %.2f%%\" % (score * 100))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/model_selection/_split.py:667: UserWarning: The least populated class in y has only 4 members, which is less than n_splits=5.\n",
            "  % (min_groups, self.n_splits)), UserWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Test score: 66.67%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "outputId": "b5eef7da-4ee2-4d7f-e9de-f2862039004c",
        "id": "kRBMmNAOdNl_",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "X, X_test, Y, y_test = train_test_split(df_news, encoded_y_news, random_state=7, test_size=0.3)\n",
        "#X, X_test, Y, y_test = train_test_split(df.values, encoded_Y, random_state=7, test_size=0.3)\n",
        "X_test, X_val, y_test, y_val = train_test_split(X_test, y_test, random_state=7, test_size=0.5)\n",
        "\n",
        "test_scores = []\n",
        "val_scores = []\n",
        "for i in range(0,10):\n",
        "\n",
        "  model = ExtraTreesClassifier(n_estimators=400, random_state=11)\n",
        "  model.fit(X, Y)\n",
        "\n",
        "  y_pred = model.predict(X_test)\n",
        "\n",
        "  score = accuracy_score(y_test, y_pred)\n",
        "  #print(\"Test score: %.2f%%\" % (score * 100))\n",
        "  test_scores.append(score * 100)\n",
        "\n",
        "  y_pred = model.predict(X_val)\n",
        "  score = accuracy_score(y_val, y_pred)\n",
        "  #print(\"Validation score: %.2f%%\" % (score * 100))\n",
        "  val_scores.append(score * 100)\n",
        "\n",
        "\n",
        "print(\"Test score: %.2f%%\" % (np.mean(test_scores)))\n",
        "print(\"Validation score: %.2f%%\" % (np.mean(val_scores)))\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test score: 73.33%\n",
            "Validation score: 66.67%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "78rziGfir8oe",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Let's manually imballance the dataset to make it look more like students dataset\n",
        "from random import randrange\n",
        "\n",
        "df_test = pd.read_csv(news_filename)\n",
        "\n",
        "safe_ids = [i for i in set([randrange(5) * i for i in range(1,40)]) if i < 100]\n",
        "df_test = df_test.drop(safe_ids)\n",
        "df_test.shape\n",
        "\n",
        "# Let's remove more samples for the first 3 classes\n",
        "def remove_by_author(author, df):\n",
        "  for idx in df.loc[df['Author'] == author].index:\n",
        "    #lets give 0.5 chance to drop each feature\n",
        "    chance = randrange(2)\n",
        "    if chance == 1:\n",
        "      df = df.drop(idx)\n",
        "  return df\n",
        "\n",
        "# Lets also give a random chance of removing these authors texts\n",
        "df_test = remove_by_author(\"'SP'\", df_test)\n",
        "df_test = remove_by_author(\"'VSAL'\", df_test)\n",
        "df_test = remove_by_author(\"'MC'\", df_test)\n",
        "df_test = remove_by_author(\"'KM'\", df_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SmKw5AODwaJP",
        "colab_type": "code",
        "outputId": "19d810a0-0b9c-4d8b-b1ea-7e2b1e88f7ab",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        }
      },
      "source": [
        "authors = {}\n",
        "for author in set(df_test['Author']):\n",
        "  authors[author] = len(df_test.loc[df_test['Author'] == author])\n",
        "\n",
        "height = authors.values()\n",
        "bars = authors.keys()\n",
        "y_pos = np.arange(len(bars))\n",
        " \n",
        "plt.bar(y_pos, height)\n",
        "plt.xticks(y_pos, bars)\n",
        " \n",
        "plt.title('Imbalanced news dataset')\n",
        "plt.ylabel('Number of samples')\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LuVq48xq4RTK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Separate data from labels\n",
        "y_news_imbalanced = df_test.pop('Author')\n",
        "\n",
        "le = LabelEncoder()\n",
        "le.fit(y_news_imbalanced)\n",
        "encoded_Y_news_imbalanced = le.transform(y_news_imbalanced)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RnYp8ZrY4Snu",
        "colab_type": "code",
        "outputId": "3f22e5cc-4306-4633-a9e5-be9bd1cd2203",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        }
      },
      "source": [
        "# Now let's use our best classifier from balanced dataset do see accuracy after imbalance\n",
        "\n",
        "X, X_test, Y, y_test = train_test_split(df_test.values, encoded_Y_news_imbalanced, random_state=7, test_size=0.3)\n",
        "X_test, X_val, y_test, y_val = train_test_split(X_test, y_test, random_state=7, test_size=0.5)\n",
        "\n",
        "# Hyper-tuning our classifier\n",
        "n_estimators = [int(x) for x in np.linspace(start = 150, stop = 1500, num = 50)]\n",
        "max_features = ['auto', 'sqrt']\n",
        "max_depth = [int(x) for x in np.linspace(10, 110, num = 11)]\n",
        "max_depth.append(None)\n",
        "min_samples_split = [2, 5, 10]\n",
        "min_samples_leaf = [1, 2, 4]\n",
        "bootstrap = [True, False]\n",
        "\n",
        "random_grid = {'n_estimators': n_estimators,\n",
        "               'max_features': max_features,\n",
        "               'max_depth': max_depth,\n",
        "               'min_samples_split': min_samples_split,\n",
        "               'min_samples_leaf': min_samples_leaf,\n",
        "               'bootstrap': bootstrap}\n",
        "                \n",
        "\n",
        "rf_random_imbalanced = RandomizedSearchCV(estimator = ExtraTreesClassifier(random_state=11), \n",
        "                                          param_distributions = random_grid, n_iter = 100, cv = 5, verbose=2, random_state=42, n_jobs = -1)\n",
        "\n",
        "rf_random_imbalanced.fit(X, Y)\n",
        "\n",
        "pprint(rf_random_imbalanced.best_params_)\n",
        "\n",
        "best_imbl_model = rf_random_imbalanced.best_estimator_\n",
        "imbl_predictions = best_imbl_model.predict(X_test)\n",
        "best_score = accuracy_score(imbl_predictions, y_test)\n",
        "print(best_score)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 5 folds for each of 100 candidates, totalling 500 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/model_selection/_split.py:667: UserWarning: The least populated class in y has only 1 members, which is less than n_splits=5.\n",
            "  % (min_groups, self.n_splits)), UserWarning)\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n",
            "[Parallel(n_jobs=-1)]: Done  37 tasks      | elapsed:   28.9s\n",
            "[Parallel(n_jobs=-1)]: Done 158 tasks      | elapsed:  2.2min\n",
            "[Parallel(n_jobs=-1)]: Done 361 tasks      | elapsed:  4.9min\n",
            "[Parallel(n_jobs=-1)]: Done 500 out of 500 | elapsed:  7.0min finished\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "{'bootstrap': False,\n",
            " 'max_depth': 60,\n",
            " 'max_features': 'sqrt',\n",
            " 'min_samples_leaf': 1,\n",
            " 'min_samples_split': 2,\n",
            " 'n_estimators': 1389}\n",
            "0.875\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WWex1pbUqdR2",
        "colab_type": "code",
        "outputId": "58d6e9cc-bd8b-40e5-956c-23afc5285fdf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "# Let's now check the accuracy of our new imbalanced model\n",
        "new_X = np.concatenate((X, X_test))\n",
        "new_Y = np.concatenate((Y, y_test))\n",
        "\n",
        "best_imbl_model.fit(new_X, new_Y)\n",
        "imbl_predictions = best_imbl_model.predict(X_val)\n",
        "best_score = accuracy_score(imbl_predictions, y_val)\n",
        "\n",
        "# It looks like class imbalance was not a problem after all\n",
        "print(best_score)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.7777777777777778\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5_GMuIacdOUA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Checking the results of a simple neural network..\n",
        "# Looks like TensorFlow 2.x is causing neural networks not to work correctly - In my machine it works :p\n",
        "X, X_test, Y, y_test = train_test_split(df_news, encoded_y_news, random_state=11, test_size=0.3)\n",
        "model = Sequential()\n",
        "model.add(Dense(12, input_shape=[54], activation='relu'))\n",
        "model.add(Dropout(0.3))\n",
        "model.add(Dense(10, activation='softmax'))\n",
        "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "\n",
        "model.fit(X, Y, epochs=250, verbose=1)\n",
        "\n",
        "scores = model.evaluate(X_test, y_test, verbose=1)\n",
        "print(\"%s: %.2f%%\" % (model.metrics_names[1], scores[1] * 100))"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}