football_tabular_fastai.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "football_tabular.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "metadata": {
        "id": "_RpcgEfOHXLK",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Tabular models"
      ]
    },
    {
      "metadata": {
        "id": "YjFZmUNuJCBE",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        " !curl https://course-v3.fast.ai/setup/colab | bash"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "tySvc_96LxPu",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "!pip install dataclasses"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "1FJyjBnAG37e",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from fastai import *\n",
        "from fastai.tabular import *"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "cvx_yC1kHXLc",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Tabular data should be in a Pandas `DataFrame`."
      ]
    },
    {
      "metadata": {
        "id": "Q1Ot6ersv7Dh",
        "colab_type": "code",
        "outputId": "e2b13cc6-c46b-42a5-a9c6-89e27866bc0d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/gdrive; to attempt to forcibly remount, call drive.mount(\"/content/gdrive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "g1sbSNL2xgm2",
        "colab_type": "code",
        "outputId": "c03750a9-6c12-43ae-9433-02bf360daaca",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "with open('/content/gdrive/My Drive/foo.txt', 'w') as f:\n",
        "  f.write('Hello Google Drive!')\n",
        "!cat /content/gdrive/My\\ Drive/foo.txt"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Hello Google Drive!"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "ViM7kLDPm4Ke",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def read_merge():\n",
        "    path = 'content/gdrive/My Drive/'\n",
        "    with open('/content/gdrive/My Drive/events.csv', 'r') as f:\n",
        "      df_events = pd.read_csv(f)\n",
        "    with open('/content/gdrive/My Drive/ginf.csv', 'r') as f:\n",
        "      df_game_info = pd.read_csv(f)\n",
        "    \n",
        "    # manually convert the dictionary.txt to python dicts\n",
        "    event_types = {1:'Attempt', 2:'Corner', 3:'Foul', 4:'Yellow card', 5:'Second yellow card', 6:'Red card', 7:'Substitution', 8:'Free kick won', 9:'Offside', 10:'Hand ball', 11:'Penalty conceded'}\n",
        "    event_types2 = {12:'Key Pass', 13:'Failed through ball', 14:'Sending off', 15:'Own goal'}\n",
        "    sides = {1:'Home', 2:'Away'}\n",
        "    shot_places = {1:'Bit too high', 2:'Blocked', 3:'Bottom left corner', 4:'Bottom right corner', 5:'Centre of the goal', 6:'High and wide', 7:'Hits the bar', 8:'Misses to the left', 9:'Misses to the right', 10:'Too high', 11:'Top centre of the goal', 12:'Top left corner', 13:'Top right corner'}\n",
        "    shot_outcomes = {1:'On target', 2:'Off target', 3:'Blocked', 4:'Hit the bar'}\n",
        "    locations = {1:'Attacking half', 2:'Defensive half', 3:'Centre of the box', 4:'Left wing', 5:'Right wing', 6:'Difficult angle and long range', 7:'Difficult angle on the left', 8:'Difficult angle on the right', 9:'Left side of the box', 10:'Left side of the six yard box', 11:'Right side of the box', 12:'Right side of the six yard box', 13:'Very close range', 14:'Penalty spot', 15:'Outside the box', 16:'Long range', 17:'More than 35 yards', 18:'More than 40 yards', 19:'Not recorded'}\n",
        "    bodyparts = {1:'right foot', 2:'left foot', 3:'head'}\n",
        "    assist_methods = {0:np.nan, 1:'Pass', 2:'Cross', 3:'Headed pass', 4:'Through ball'}\n",
        "    situations = {1:'Open play', 2:'Set piece', 3:'Corner', 4:'Free kick'}\n",
        "    \n",
        "    # Mapping the dicts into the events dataframe\n",
        "    df_events['event_type'] =   df_events['event_type'].map(event_types)\n",
        "    df_events['event_type2'] =  df_events['event_type2'].map(event_types2)\n",
        "    df_events['side'] =         df_events['side'].map(sides)\n",
        "    df_events['shot_place'] =   df_events['shot_place'].map(shot_places)\n",
        "    df_events['shot_outcome']=  df_events['shot_outcome'].map(shot_outcomes)\n",
        "    df_events['location'] =     df_events['location'].map(locations)\n",
        "    df_events['bodypart'] =     df_events['bodypart'].map(bodyparts)\n",
        "    df_events['assist_method']= df_events['assist_method'].map(assist_methods)\n",
        "    df_events['situation'] =    df_events['situation'].map(situations)\n",
        "\n",
        "    ## Naming the leagues with their popular names, which will make thinks much clear for us\n",
        "    leagues = {'E0': 'Premier League', 'SP1': 'La Liga',\n",
        "              'I1': 'Serie A', 'F1': 'League One', 'D1': 'Bundesliga'}\n",
        "\n",
        "    ## Map them to events\n",
        "    df_game_info.league = df_game_info.league.map(leagues)\n",
        "\n",
        "    # Merge into one table (merge other dataset to have country, league, date and season)\n",
        "    df_events = df_events.merge(df_game_info ,how = 'left')\n",
        "    \n",
        "    return df_events\n",
        "  \n",
        "def fill_unk(df):\n",
        "    '''\n",
        "    Fill the following categorical features in df with 'UNK':\n",
        "    shot_place, player, shot_outcome, bodypart, location and assist_method\n",
        "    '''\n",
        "    # Fill the required features with new class 'unknown'\n",
        "    df.shot_place.fillna('UNK'   ,inplace = True)\n",
        "    df.player.fillna('UNK'       ,inplace = True)\n",
        "    df.shot_outcome.fillna('UNK' ,inplace = True)\n",
        "    df.bodypart.fillna('UNK'     ,inplace = True)\n",
        "    df.location.fillna('UNK'     ,inplace = True)\n",
        "    df.assist_method.fillna('UNK',inplace = True)\n",
        "    df.situation.fillna('UNK'    ,inplace = True)\n",
        "    df.location.replace('Not recorded', 'UNK', inplace= True)\n",
        "    \n",
        "def fix_dtypes(df, cats):\n",
        "  '''\n",
        "  fixes the category datatypes for cats features\n",
        "  '''\n",
        "  # Notice that a lot of the floats and ints are in fact categoricals\n",
        "  # We can fix this with Pandas' astype function\n",
        "  d = dict.fromkeys(cats,'category')\n",
        "  df = df.astype(d)\n",
        "  df['is_goal'] = df['is_goal'].astype('bool') # this is a bool\n",
        "  df['fast_break'] = df['fast_break'].astype('bool')\n",
        "\n",
        "  df.odd_a = df.odd_a.astype('float16')\n",
        "  df.odd_d = df.odd_d.astype('float16')\n",
        "  df.odd_h = df.odd_h.astype('float16')\n",
        "  df.odd_over  = df.odd_over.astype('float16')\n",
        "  df.odd_under = df.odd_under.astype('float16')               \n",
        "  df.odd_bts   = df.odd_bts.astype('float16')\n",
        "  df.odd_bts_n = df.odd_bts_n.astype('float16')\n",
        "\n",
        "  df.ftag = df.ftag.astype('int8')\n",
        "  df.fthg = df.fthg.astype('int8')\n",
        "  df.time = df.time.astype('int8')\n",
        "\n",
        "  df.season =     df.season.astype('category')\n",
        "  df.league =     df.league.astype('category')\n",
        "  df.country =    df.country.astype('category')\n",
        "  df.date = pd.to_datetime(df.date)\n",
        "  return df\n",
        "\n",
        "def missing_values_table(df):\n",
        "  '''\n",
        "  Function to calculate missing values by column# Funct \n",
        "  '''\n",
        "  # Total missing values\n",
        "  mis_val = df.isnull().sum()\n",
        "\n",
        "  # Percentage of missing values\n",
        "  mis_val_percent = 100 * df.isnull().sum() / len(df)\n",
        "\n",
        "  # Make a table with the results\n",
        "  mis_val_table = pd.concat([mis_val, mis_val_percent], axis=1)\n",
        "\n",
        "  # Rename the columns\n",
        "  mis_val_table_ren_columns = mis_val_table.rename(\n",
        "  columns = {0 : 'Missing Values', 1 : '% of Total Values'})\n",
        "\n",
        "  # Sort the table by percentage of missing descending\n",
        "  mis_val_table_ren_columns = mis_val_table_ren_columns[\n",
        "      mis_val_table_ren_columns.iloc[:,1] != 0].sort_values(\n",
        "  '% of Total Values', ascending=False).round(1)\n",
        "\n",
        "  # Print some summary information\n",
        "  print (\"Your selected dataframe has \" + str(df.shape[1]) + \" columns.\\n\"      \n",
        "      \"There are \" + str(mis_val_table_ren_columns.shape[0]) +\n",
        "        \" columns that have missing values.\")\n",
        "\n",
        "  # Return the dataframe with missing information\n",
        "  return mis_val_table_ren_columns\n",
        "\n",
        "def select_feats(df, feat_cols):\n",
        "  '''\n",
        "  Manually choose features (feat_cols) from a dataframe (df)\n",
        "  '''\n",
        "  return df[feat_cols]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "HT3tLDKZm4tg",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Read the dataset \n",
        "df = read_merge()\n",
        "\n",
        "# Fill the null values with a new category 'UNK' for 'Unknown'\n",
        "# fill_unk(df)\n",
        "\n",
        "\n",
        "# Fix the datatypes of some categorical columns\n",
        "categoricals = ['id_odsp', 'id_event', 'event_type','event_type2', 'player2', 'event_team', 'opponent', 'shot_place', 'shot_outcome', 'location', 'bodypart', 'assist_method', 'situation', 'side', 'player']\n",
        "df = fix_dtypes(df, categoricals)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "2co7HUu7eO49",
        "colab_type": "code",
        "outputId": "99b98840-f0bf-44c7-e6d7-a78a0ea37804",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 124
        }
      },
      "cell_type": "code",
      "source": [
        "feats = ['odd_h', 'odd_d', 'odd_a', 'event_team', 'player', 'league',\n",
        "             'assist_method', 'location', \n",
        "             'side', 'shot_place', 'situation', \n",
        "             'bodypart', 'time', 'is_goal']\n",
        "\n",
        "df_feats = select_feats(df, feats)\n",
        "df_feats['first_half'] = df_feats.time <= 45"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:7: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame.\n",
            "Try using .loc[row_indexer,col_indexer] = value instead\n",
            "\n",
            "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
            "  import sys\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "metadata": {
        "id": "Nx71VYMfm5Lc",
        "colab_type": "code",
        "outputId": "c53ddefa-28bc-41e4-bdc8-59fdc3f20a96",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 374
        }
      },
      "cell_type": "code",
      "source": [
        "# df_feats.dropna(inplace =True)\n",
        "df_feats.info()"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 941009 entries, 0 to 941008\n",
            "Data columns (total 15 columns):\n",
            "odd_h            941009 non-null float16\n",
            "odd_d            941009 non-null float16\n",
            "odd_a            941009 non-null float16\n",
            "event_team       941009 non-null category\n",
            "player           880009 non-null category\n",
            "league           941009 non-null category\n",
            "assist_method    167905 non-null category\n",
            "location         467067 non-null category\n",
            "side             941009 non-null category\n",
            "shot_place       227459 non-null category\n",
            "situation        229137 non-null category\n",
            "bodypart         229185 non-null category\n",
            "time             941009 non-null int8\n",
            "is_goal          941009 non-null bool\n",
            "first_half       941009 non-null bool\n",
            "dtypes: bool(2), category(9), float16(3), int8(1)\n",
            "memory usage: 25.3 MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "dARHKufDMqHM",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "training_features = df_feats.drop(columns=['is_goal', 'time'])\n",
        "training_target = df_feats.is_goal\n",
        "\n",
        "x_train, x_test, y_train, y_test = train_test_split(training_features, training_target,\n",
        "                                                  test_size = .1,\n",
        "                                                  random_state=12)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "gIogBeT5m5tk",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Final chosen/engineered features\n",
        "feats_cat = ['assist_method', 'location', 'event_team', 'player', 'league',\n",
        "             'side', 'shot_place', 'situation', \n",
        "             'bodypart', 'first_half']\n",
        "feats_num = ['odd_h', 'odd_d', 'odd_a']\n",
        "target_var = 'is_goal'\n",
        "procs = [FillMissing, Categorify, Normalize]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "g7TjUm2tDaDz",
        "colab_type": "code",
        "outputId": "463fc955-56b5-4776-afe8-59f9fac13e50",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 178
        }
      },
      "cell_type": "code",
      "source": [
        "!pip3 install scikit-learn\n",
        "!pip3 install -U imbalanced-learn\n",
        "!pip3 install scipy"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: scikit-learn in /usr/local/lib/python3.6/dist-packages (0.20.1)\n",
            "Requirement already satisfied: scipy>=0.13.3 in /usr/local/lib/python3.6/dist-packages (from scikit-learn) (1.1.0)\n",
            "Requirement already satisfied: numpy>=1.8.2 in /usr/local/lib/python3.6/dist-packages (from scikit-learn) (1.14.6)\n",
            "Requirement already up-to-date: imbalanced-learn in /usr/local/lib/python3.6/dist-packages (0.4.3)\n",
            "Requirement already satisfied, skipping upgrade: scikit-learn>=0.20 in /usr/local/lib/python3.6/dist-packages (from imbalanced-learn) (0.20.1)\n",
            "Requirement already satisfied, skipping upgrade: numpy>=1.8.2 in /usr/local/lib/python3.6/dist-packages (from imbalanced-learn) (1.14.6)\n",
            "Requirement already satisfied, skipping upgrade: scipy>=0.13.3 in /usr/local/lib/python3.6/dist-packages (from imbalanced-learn) (1.1.0)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.6/dist-packages (1.1.0)\n",
            "Requirement already satisfied: numpy>=1.8.2 in /usr/local/lib/python3.6/dist-packages (from scipy) (1.14.6)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "PMbH4YaNC_KL",
        "colab_type": "code",
        "outputId": "c0484019-c0b4-4352-8207-bec233d06f49",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "cell_type": "code",
      "source": [
        "from imblearn.under_sampling import RandomUnderSampler\n",
        "from collections import Counter\n",
        "\n",
        "print('Original dataset shape %s' % Counter(df_feats.is_goal))\n",
        "\n",
        "rus = RandomUnderSampler(random_state=42)\n",
        "X_res_u, y_res_u = rus.fit_resample(x_train, y_train)\n",
        "\n",
        "print('Original dataset shape %s' % Counter(y_res_u))"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Original dataset shape Counter({False: 916563, True: 24446})\n",
            "Original dataset shape Counter({False: 21976, True: 21976})\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "_se8_AfJm6Ch",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# To split our data into training and validation sets, we use valid indexes\n",
        "\n",
        "valid_idx = range(len(X_res_u)-2000, len(X_res_u))\n",
        "\n",
        "path = '/content/gdrive/My Drive/'"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "dHvM_6aaYDur",
        "colab_type": "code",
        "outputId": "6a2b32cf-f08d-48ca-d69b-aaadcabfc97a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "valid_idx\n",
        "# X_res_u.shape"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "range(41952, 43952)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "metadata": {
        "id": "otzcOgMaSgkK",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "both_arr = np.hstack((X_res_u, y_res_u.reshape(-1,1)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "gow6oGOPStY1",
        "colab_type": "code",
        "outputId": "2b29bd71-12f4-452f-98ba-22305496ab5a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        }
      },
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "cols_all = ['odd_h', 'odd_d', 'odd_a', 'event_team', 'player', 'league',\n",
        "         'assist_method', 'location', \n",
        "         'side', 'shot_place', 'situation', \n",
        "         'bodypart', 'first_half', 'is_goal']\n",
        "df_train = pd.DataFrame(data = both_arr, columns = cols_all)\n",
        "procs=[FillMissing, Categorify, Normalize]\n",
        "df_train.head()\n"
      ],
      "execution_count": 15,
      "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>odd_h</th>\n",
              "      <th>odd_d</th>\n",
              "      <th>odd_a</th>\n",
              "      <th>event_team</th>\n",
              "      <th>player</th>\n",
              "      <th>league</th>\n",
              "      <th>assist_method</th>\n",
              "      <th>location</th>\n",
              "      <th>side</th>\n",
              "      <th>shot_place</th>\n",
              "      <th>situation</th>\n",
              "      <th>bodypart</th>\n",
              "      <th>first_half</th>\n",
              "      <th>is_goal</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1.94043</td>\n",
              "      <td>3.59961</td>\n",
              "      <td>5.19922</td>\n",
              "      <td>Atalanta</td>\n",
              "      <td>luca cigarini</td>\n",
              "      <td>Serie A</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Away</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>17.7812</td>\n",
              "      <td>7.80078</td>\n",
              "      <td>1.25</td>\n",
              "      <td>Real Valladolid</td>\n",
              "      <td>fausto rossi</td>\n",
              "      <td>La Liga</td>\n",
              "      <td>Headed pass</td>\n",
              "      <td>Outside the box</td>\n",
              "      <td>Home</td>\n",
              "      <td>Blocked</td>\n",
              "      <td>Open play</td>\n",
              "      <td>right foot</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2.09961</td>\n",
              "      <td>3.5</td>\n",
              "      <td>5.10156</td>\n",
              "      <td>Evian Thonon Gaillard</td>\n",
              "      <td>nicki bille</td>\n",
              "      <td>League One</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Away</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>True</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2.66992</td>\n",
              "      <td>3.38086</td>\n",
              "      <td>3.15039</td>\n",
              "      <td>Osasuna</td>\n",
              "      <td>de las cuevas</td>\n",
              "      <td>La Liga</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Home</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2.69922</td>\n",
              "      <td>3.25</td>\n",
              "      <td>3.25</td>\n",
              "      <td>AC Ajaccio</td>\n",
              "      <td>grenddy perozo</td>\n",
              "      <td>League One</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Home</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "     odd_h    odd_d    odd_a             event_team          player  \\\n",
              "0  1.94043  3.59961  5.19922               Atalanta   luca cigarini   \n",
              "1  17.7812  7.80078     1.25        Real Valladolid    fausto rossi   \n",
              "2  2.09961      3.5  5.10156  Evian Thonon Gaillard     nicki bille   \n",
              "3  2.66992  3.38086  3.15039                Osasuna   de las cuevas   \n",
              "4  2.69922     3.25     3.25             AC Ajaccio  grenddy perozo   \n",
              "\n",
              "       league assist_method         location  side shot_place  situation  \\\n",
              "0     Serie A           NaN              NaN  Away        NaN        NaN   \n",
              "1     La Liga   Headed pass  Outside the box  Home    Blocked  Open play   \n",
              "2  League One           NaN              NaN  Away        NaN        NaN   \n",
              "3     La Liga           NaN              NaN  Home        NaN        NaN   \n",
              "4  League One           NaN              NaN  Home        NaN        NaN   \n",
              "\n",
              "     bodypart first_half is_goal  \n",
              "0         NaN       True   False  \n",
              "1  right foot       True   False  \n",
              "2         NaN       True   False  \n",
              "3         NaN      False   False  \n",
              "4         NaN      False   False  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "metadata": {
        "id": "7HsxQI-LZQUV",
        "colab_type": "code",
        "outputId": "1996a119-11c3-4171-d970-32d030332fd4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        }
      },
      "cell_type": "code",
      "source": [
        "d_num = dict.fromkeys(feats_num, 'float32')\n",
        "df_train = df_train.astype(d_num)\n",
        "\n",
        "d_cat = dict.fromkeys(feats_cat, 'category')\n",
        "df_train = df_train.astype(d_cat)\n",
        "\n",
        "df_train['is_goal'] = df_train['is_goal'].astype('bool') # this is a bool\n",
        "df_train.dtypes"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "odd_h             float32\n",
              "odd_d             float32\n",
              "odd_a             float32\n",
              "event_team       category\n",
              "player           category\n",
              "league           category\n",
              "assist_method    category\n",
              "location         category\n",
              "side             category\n",
              "shot_place       category\n",
              "situation        category\n",
              "bodypart         category\n",
              "first_half       category\n",
              "is_goal              bool\n",
              "dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "metadata": {
        "id": "VHAKg3blO8rw",
        "colab_type": "code",
        "outputId": "3bbf07db-4fd7-4a24-bb53-a9db38b0fc5f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "test = TabularList.from_df(df_train.iloc[valid_idx].copy(), cat_names = feats_cat, cont_names= feats_num)\n",
        "\n",
        "data = (TabularList.from_df(df_train, cat_names = feats_cat, cont_names= feats_num)\n",
        "                   .split_by_idx(valid_idx)\n",
        "                   .label_from_df(cols=target_var)\n",
        "                   .add_test(test)        \n",
        "                   .databunch())\n",
        "\n",
        "\n",
        "\n",
        "# data = TabularDataBunch.from_df(path, df_train, target_var, valid_idx=valid_idx, cat_names=feats_cat)\n",
        "print(data.train_ds.cont_names)  # `cont_names` defaults to: set(df)-set(cat_names)-{dep_var}"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['odd_h', 'odd_d', 'odd_a']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "zEy8uEZxPTRi",
        "colab_type": "code",
        "outputId": "e23e2d8b-1df3-43e3-84e3-4448cc3b991d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 254
        }
      },
      "cell_type": "code",
      "source": [
        "data.show_batch(rows=10)"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<table>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <col width='10%'>  <tr>\n",
              "    <th>assist_method</th>\n",
              "    <th>location</th>\n",
              "    <th>event_team</th>\n",
              "    <th>player</th>\n",
              "    <th>league</th>\n",
              "    <th>side</th>\n",
              "    <th>shot_place</th>\n",
              "    <th>situation</th>\n",
              "    <th>bodypart</th>\n",
              "    <th>first_half</th>\n",
              "    <th>odd_h</th>\n",
              "    <th>odd_d</th>\n",
              "    <th>odd_a</th>\n",
              "    <th>target</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>Guingamp</th>\n",
              "    <th>lars jacobsen</th>\n",
              "    <th>League One</th>\n",
              "    <th>Away</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>False</th>\n",
              "    <th>2.8008</th>\n",
              "    <th>3.1309</th>\n",
              "    <th>3.2598</th>\n",
              "    <th>False</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>Cross</th>\n",
              "    <th>Centre of the box</th>\n",
              "    <th>Atalanta</th>\n",
              "    <th>cristian raimondi</th>\n",
              "    <th>Serie A</th>\n",
              "    <th>Home</th>\n",
              "    <th>Bottom right corner</th>\n",
              "    <th>Open play</th>\n",
              "    <th>head</th>\n",
              "    <th>False</th>\n",
              "    <th>2.9297</th>\n",
              "    <th>2.4805</th>\n",
              "    <th>4.6484</th>\n",
              "    <th>False</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>Pass</th>\n",
              "    <th>Centre of the box</th>\n",
              "    <th>AS Monaco</th>\n",
              "    <th>lacina traore</th>\n",
              "    <th>League One</th>\n",
              "    <th>Away</th>\n",
              "    <th>Bottom right corner</th>\n",
              "    <th>Open play</th>\n",
              "    <th>left foot</th>\n",
              "    <th>False</th>\n",
              "    <th>4.2500</th>\n",
              "    <th>3.3398</th>\n",
              "    <th>2.3398</th>\n",
              "    <th>True</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>Through ball</th>\n",
              "    <th>Left side of the box</th>\n",
              "    <th>Everton</th>\n",
              "    <th>arouna kone</th>\n",
              "    <th>Premier League</th>\n",
              "    <th>Home</th>\n",
              "    <th>Top right corner</th>\n",
              "    <th>Open play</th>\n",
              "    <th>left foot</th>\n",
              "    <th>False</th>\n",
              "    <th>1.5498</th>\n",
              "    <th>4.8008</th>\n",
              "    <th>7.8008</th>\n",
              "    <th>True</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>#na#</th>\n",
              "    <th>Defensive half</th>\n",
              "    <th>Nantes</th>\n",
              "    <th>itay shechter</th>\n",
              "    <th>League One</th>\n",
              "    <th>Away</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>True</th>\n",
              "    <th>1.5996</th>\n",
              "    <th>4.0508</th>\n",
              "    <th>8.6016</th>\n",
              "    <th>False</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>#na#</th>\n",
              "    <th>Defensive half</th>\n",
              "    <th>Stade Rennes</th>\n",
              "    <th>georges mandjeck</th>\n",
              "    <th>League One</th>\n",
              "    <th>Away</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>True</th>\n",
              "    <th>1.7197</th>\n",
              "    <th>3.8008</th>\n",
              "    <th>6.1484</th>\n",
              "    <th>False</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>Juventus</th>\n",
              "    <th>matteo rubin</th>\n",
              "    <th>Serie A</th>\n",
              "    <th>Away</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>True</th>\n",
              "    <th>5.8984</th>\n",
              "    <th>3.7598</th>\n",
              "    <th>1.7695</th>\n",
              "    <th>False</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>Eibar</th>\n",
              "    <th>mauro dos santos</th>\n",
              "    <th>La Liga</th>\n",
              "    <th>Away</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>False</th>\n",
              "    <th>2.1191</th>\n",
              "    <th>3.6992</th>\n",
              "    <th>3.9004</th>\n",
              "    <th>False</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>Sevilla</th>\n",
              "    <th>vicente iborra</th>\n",
              "    <th>La Liga</th>\n",
              "    <th>Away</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>#na#</th>\n",
              "    <th>False</th>\n",
              "    <th>3.5000</th>\n",
              "    <th>3.8008</th>\n",
              "    <th>2.2598</th>\n",
              "    <th>False</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>Through ball</th>\n",
              "    <th>Left side of the box</th>\n",
              "    <th>Torino</th>\n",
              "    <th>riccardo meggiorini</th>\n",
              "    <th>Serie A</th>\n",
              "    <th>Home</th>\n",
              "    <th>Misses to the left</th>\n",
              "    <th>Open play</th>\n",
              "    <th>left foot</th>\n",
              "    <th>False</th>\n",
              "    <th>3.0508</th>\n",
              "    <th>3.3594</th>\n",
              "    <th>2.7402</th>\n",
              "    <th>False</th>\n",
              "  </tr>\n",
              "</table>\n"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "VvtFJ4BjPT1-",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "learn = tabular_learner(data, layers=[200,100], metrics=accuracy)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "uNlEFAuyO9m1",
        "colab_type": "code",
        "outputId": "bc8b5477-69ca-4a00-d2e0-a5106c2ed9d7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 453
        }
      },
      "cell_type": "code",
      "source": [
        "# find best lr\n",
        "learn.lr_find()\n",
        "\n",
        "# find best lr\n",
        "learn.recorder.plot()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "    <div>\n",
              "        <style>\n",
              "            /* Turns off some styling */\n",
              "            progress {\n",
              "                /* gets rid of default border in Firefox and Opera. */\n",
              "                border: none;\n",
              "                /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "                background-size: auto;\n",
              "            }\n",
              "            .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "                background: #F44336;\n",
              "            }\n",
              "        </style>\n",
              "      <progress value='0' class='' max='1', style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
              "      0.00% [0/1 00:00<00:00]\n",
              "    </div>\n",
              "    \n",
              "<table style='width:300px; margin-bottom:10px'>\n",
              "  <tr>\n",
              "    <th>epoch</th>\n",
              "    <th>train_loss</th>\n",
              "    <th>valid_loss</th>\n",
              "    <th>accuracy</th>\n",
              "  </tr>\n",
              "</table>\n",
              "\n",
              "\n",
              "    <div>\n",
              "        <style>\n",
              "            /* Turns off some styling */\n",
              "            progress {\n",
              "                /* gets rid of default border in Firefox and Opera. */\n",
              "                border: none;\n",
              "                /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "                background-size: auto;\n",
              "            }\n",
              "            .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "                background: #F44336;\n",
              "            }\n",
              "        </style>\n",
              "      <progress value='0' class='progress-bar-interrupted' max='656', style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
              "      Interrupted\n",
              "    </div>\n",
              "    "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfcAAAFYCAYAAABOP7UcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nwVwAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7ffa68222550>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "eKmk07MgHXL4",
        "colab_type": "code",
        "outputId": "1f83c3aa-2735-4719-e225-f08da1c6b1ce",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 287
        }
      },
      "cell_type": "code",
      "source": [
        "learn.fit(10, 1e-1)\n"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "Total time: 02:01 <p><table style='width:300px; margin-bottom:10px'>\n",
              "  <tr>\n",
              "    <th>epoch</th>\n",
              "    <th>train_loss</th>\n",
              "    <th>valid_loss</th>\n",
              "    <th>accuracy</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>1</th>\n",
              "    <th>0.112653</th>\n",
              "    <th>0.010898</th>\n",
              "    <th>0.999000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>2</th>\n",
              "    <th>0.151258</th>\n",
              "    <th>0.122586</th>\n",
              "    <th>0.963500</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>3</th>\n",
              "    <th>0.239442</th>\n",
              "    <th>0.006953</th>\n",
              "    <th>1.000000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>4</th>\n",
              "    <th>0.114339</th>\n",
              "    <th>0.222713</th>\n",
              "    <th>0.917500</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>5</th>\n",
              "    <th>0.100520</th>\n",
              "    <th>0.004154</th>\n",
              "    <th>1.000000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>6</th>\n",
              "    <th>0.107814</th>\n",
              "    <th>0.100504</th>\n",
              "    <th>0.986000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>7</th>\n",
              "    <th>0.104141</th>\n",
              "    <th>0.386878</th>\n",
              "    <th>0.802000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>8</th>\n",
              "    <th>0.105179</th>\n",
              "    <th>0.025666</th>\n",
              "    <th>0.999500</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>9</th>\n",
              "    <th>0.103425</th>\n",
              "    <th>0.135911</th>\n",
              "    <th>0.961000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>10</th>\n",
              "    <th>0.115187</th>\n",
              "    <th>0.044808</th>\n",
              "    <th>0.994000</th>\n",
              "  </tr>\n",
              "</table>\n"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "vQAZsLUlTDOM",
        "colab_type": "code",
        "outputId": "0a0cc71e-132c-4c35-e9a2-08822b207d2e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "learn.validate()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.10538929, tensor(0.9820)]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 120
        }
      ]
    },
    {
      "metadata": {
        "id": "2FFiUEpbHXMY",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Inference"
      ]
    },
    {
      "metadata": {
        "id": "R8sL8BdZHXMd",
        "colab_type": "code",
        "outputId": "56c9e358-dac1-43d4-8dff-78be92eda5c7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "row = df_train.iloc[0]\n",
        "# print(df_train.dtypes)\n",
        "# print(row)\n",
        "type(row)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "pandas.core.series.Series"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 53
        }
      ]
    },
    {
      "metadata": {
        "id": "2Hbcm8acHXMj",
        "colab_type": "code",
        "outputId": "533b67d2-b38c-46bd-b420-abc58be7d717",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 142
        }
      },
      "cell_type": "code",
      "source": [
        "learn.get_preds()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[tensor([[0.0143, 0.9857],\n",
              "         [0.0122, 0.9878],\n",
              "         [0.0712, 0.9288],\n",
              "         ...,\n",
              "         [0.0365, 0.9635],\n",
              "         [0.1147, 0.8853],\n",
              "         [0.1013, 0.8987]]), tensor([1, 1, 1,  ..., 1, 1, 1])]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 116
        }
      ]
    },
    {
      "metadata": {
        "id": "ia9M5uDueOIO",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Predicting the result according to Odds"
      ]
    },
    {
      "metadata": {
        "id": "houQw7n6eSi5",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "90e5d27b-dcba-4030-ed7f-b570dcceb262"
      },
      "cell_type": "code",
      "source": [
        "def odds_pred_result(odd_h,odd_d,odd_a):\n",
        "    '''\n",
        "    Predict game final result according to the odds for (home win, draw, and )\n",
        "    Predict result of a match based on the highest odds for each result for a particular match\n",
        "    '''\n",
        "    if odd_h<odd_d and odd_h<odd_a:\n",
        "        return(\"Home Win\")\n",
        "    elif odd_d<odd_h and odd_d<odd_a:\n",
        "        return(\"Draw\")\n",
        "    elif odd_a<odd_d and odd_a<odd_h:\n",
        "        return(\"Away Win\")\n",
        "    \n",
        "\n",
        "\n",
        "# The next function would return the actual result of the match.\n",
        "def actual_result(fthg,ftag):\n",
        "    if fthg>ftag:\n",
        "        return(\"Home Win\")\n",
        "    elif fthg==ftag:\n",
        "        return(\"Draw\")\n",
        "    elif fthg<ftag:\n",
        "        return(\"Away Win\")\n",
        "    \n",
        "def actual_result_encode(fthg,ftag):\n",
        "    if fthg>ftag:\n",
        "        return (1)\n",
        "    elif fthg==ftag:\n",
        "        return (2)\n",
        "    elif fthg<ftag:\n",
        "        return (3)\n",
        "      \n",
        "# We now compare the chances of us predicting the result of the game correctly just by looking at the odds.\n",
        "def check_pred(data):\n",
        "    correct=0\n",
        "    wrong=0\n",
        "    for i in range(1,data.shape[0]+1):\n",
        "        odd_h=data[i-1:i][\"odd_h\"].item()\n",
        "        odd_d=data[i-1:i][\"odd_d\"].item()\n",
        "        odd_a=data[i-1:i][\"odd_a\"].item()\n",
        "        fthg=data[i-1:i][\"fthg\"].item()\n",
        "        ftag=data[i-1:i][\"ftag\"].item()\n",
        "        oddsresult=odds_pred_result(odd_h,odd_d,odd_a)\n",
        "        actresult=actual_result(fthg,ftag)\n",
        "        if oddsresult==actresult:\n",
        "            correct+=1\n",
        "        else:\n",
        "            wrong+=1\n",
        "    return(str(\"%.2f\"%(correct/(correct+wrong)))+str(\"% correct\"))\n",
        "\n",
        "with open('/content/gdrive/My Drive/ginf.csv', 'r') as f:\n",
        "  df_game_info = pd.read_csv(f)\n",
        "  \n",
        "with open('/content/gdrive/My Drive/events.csv', 'r') as f:\n",
        "  df_events = pd.read_csv(f)\n",
        "  \n",
        "  \n",
        "print('Initial guess: ', check_pred(df_game_info))\n"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'0.53% correct'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 23
        }
      ]
    },
    {
      "metadata": {
        "id": "j7Xy0Y8vQ4W6",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "71916b1a-304d-4b10-ec04-5f823d92014d"
      },
      "cell_type": "code",
      "source": [
        "# x_odds = df_game_info.iloc[:,9:14]\n",
        "\n",
        "     \n",
        "df_odds_all = df_game_info.join(df_events, lsuffix='_odds', rsuffix='_events')# , how = 'inner')#, left_on = 'id_odsp', right_on = 'id_odsp')\n",
        "\n",
        "df_odds_all = select_feats(df_odds_all, ['odd_h', 'odd_d', 'odd_a', 'event_team', 'opponent', 'league', 'side', 'fthg', 'ftag'])\n",
        "\n",
        "\n",
        "# manually convert the dictionary.txt to python dicts\n",
        "sides = {1:'Home', 2:'Away'}\n",
        "leagues = {'E0': 'Premier League', 'SP1': 'La Liga',\n",
        "              'I1': 'Serie A', 'F1': 'League One', 'D1': 'Bundesliga'}\n",
        "\n",
        "    \n",
        "# Mapping the dicts into the events dataframe\n",
        "df_odds_all['side'] = df_odds_all['side'].map(sides)\n",
        "df_odds_all.league = df_odds_all.league.map(leagues)\n",
        "    \n",
        "    \n",
        "# df_odds_all.info()\n",
        "\n",
        "result=[]\n",
        "for i in range(1, df_odds_all.shape[0]+1):\n",
        "    result.append(actual_result_encode(df_odds_all[i-1:i][\"fthg\"].item(),\n",
        "                                       df_odds_all[i-1:i][\"ftag\"].item()))\n",
        "    \n",
        "df_odds_all['result'] = pd.DataFrame(result)\n",
        "\n",
        "df_odds_all[\"diff_h_d\"] = abs(df_odds_all[\"odd_h\"] - df_odds_all[\"odd_d\"])\n",
        "df_odds_all[\"diff_d_a\"] = abs(df_odds_all[\"odd_d\"] - df_odds_all[\"odd_a\"])\n",
        "df_odds_all[\"diff_h_a\"] = abs(df_odds_all[\"odd_h\"] - df_odds_all[\"odd_a\"])\n",
        "\n",
        "df_odds_all = df_odds_all.drop([\"fthg\",\"ftag\"],axis=1)\n",
        "\n",
        "df_odds_all.tail()"
      ],
      "execution_count": 56,
      "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>odd_h</th>\n",
              "      <th>odd_d</th>\n",
              "      <th>odd_a</th>\n",
              "      <th>event_team</th>\n",
              "      <th>opponent</th>\n",
              "      <th>league</th>\n",
              "      <th>side</th>\n",
              "      <th>result</th>\n",
              "      <th>diff_h_d</th>\n",
              "      <th>diff_d_a</th>\n",
              "      <th>diff_h_a</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>10107</th>\n",
              "      <td>1.97</td>\n",
              "      <td>4.35</td>\n",
              "      <td>8.0</td>\n",
              "      <td>Lorient</td>\n",
              "      <td>Sochaux</td>\n",
              "      <td>Serie A</td>\n",
              "      <td>Away</td>\n",
              "      <td>2</td>\n",
              "      <td>2.38</td>\n",
              "      <td>3.65</td>\n",
              "      <td>6.03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10108</th>\n",
              "      <td>1.19</td>\n",
              "      <td>8.50</td>\n",
              "      <td>20.0</td>\n",
              "      <td>Sochaux</td>\n",
              "      <td>Lorient</td>\n",
              "      <td>Premier League</td>\n",
              "      <td>Home</td>\n",
              "      <td>1</td>\n",
              "      <td>7.31</td>\n",
              "      <td>11.50</td>\n",
              "      <td>18.81</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10109</th>\n",
              "      <td>1.32</td>\n",
              "      <td>6.24</td>\n",
              "      <td>11.5</td>\n",
              "      <td>Lorient</td>\n",
              "      <td>Sochaux</td>\n",
              "      <td>League One</td>\n",
              "      <td>Away</td>\n",
              "      <td>1</td>\n",
              "      <td>4.92</td>\n",
              "      <td>5.26</td>\n",
              "      <td>10.18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10110</th>\n",
              "      <td>1.74</td>\n",
              "      <td>4.07</td>\n",
              "      <td>5.9</td>\n",
              "      <td>Sochaux</td>\n",
              "      <td>Lorient</td>\n",
              "      <td>La Liga</td>\n",
              "      <td>Home</td>\n",
              "      <td>2</td>\n",
              "      <td>2.33</td>\n",
              "      <td>1.83</td>\n",
              "      <td>4.16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10111</th>\n",
              "      <td>1.79</td>\n",
              "      <td>3.96</td>\n",
              "      <td>5.4</td>\n",
              "      <td>Lorient</td>\n",
              "      <td>Sochaux</td>\n",
              "      <td>Serie A</td>\n",
              "      <td>Away</td>\n",
              "      <td>1</td>\n",
              "      <td>2.17</td>\n",
              "      <td>1.44</td>\n",
              "      <td>3.61</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "       odd_h  odd_d  odd_a event_team opponent          league  side  result  \\\n",
              "10107   1.97   4.35    8.0    Lorient  Sochaux         Serie A  Away       2   \n",
              "10108   1.19   8.50   20.0    Sochaux  Lorient  Premier League  Home       1   \n",
              "10109   1.32   6.24   11.5    Lorient  Sochaux      League One  Away       1   \n",
              "10110   1.74   4.07    5.9    Sochaux  Lorient         La Liga  Home       2   \n",
              "10111   1.79   3.96    5.4    Lorient  Sochaux         Serie A  Away       1   \n",
              "\n",
              "       diff_h_d  diff_d_a  diff_h_a  \n",
              "10107      2.38      3.65      6.03  \n",
              "10108      7.31     11.50     18.81  \n",
              "10109      4.92      5.26     10.18  \n",
              "10110      2.33      1.83      4.16  \n",
              "10111      2.17      1.44      3.61  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        }
      ]
    },
    {
      "metadata": {
        "id": "FkNdQvZhesak",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 303
        },
        "outputId": "5c5c0436-eee8-4682-ff06-25ce14dd3eb0"
      },
      "cell_type": "code",
      "source": [
        "x_odds_train, x_odds_test, y_odds_train, y_odds_test = train_test_split(df_odds_all.drop(['result'], axis=1), df_odds_all.result, test_size=0.2,random_state=0)\n",
        "\n",
        "\n",
        "df_odds_train = x_odds_train.copy()\n",
        "df_odds_train['result'] = y_odds_train\n",
        "df_odds_train.result = df_odds_train.result.astype('category')\n",
        "\n",
        "df_odds_train.info()\n"
      ],
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 8089 entries, 8556 to 2732\n",
            "Data columns (total 11 columns):\n",
            "odd_h         8089 non-null float64\n",
            "odd_d         8089 non-null float64\n",
            "odd_a         8089 non-null float64\n",
            "event_team    8089 non-null object\n",
            "opponent      8089 non-null object\n",
            "league        8089 non-null object\n",
            "side          8089 non-null object\n",
            "diff_h_d      8089 non-null float64\n",
            "diff_d_a      8089 non-null float64\n",
            "diff_h_a      8089 non-null float64\n",
            "result        8089 non-null category\n",
            "dtypes: category(1), float64(6), object(4)\n",
            "memory usage: 703.1+ KB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "BalRxhi1RYcb",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 321
        },
        "outputId": "8777e761-ef43-4649-81d2-3151a6d81dc3"
      },
      "cell_type": "code",
      "source": [
        "valid_idx_odds = range(len(df_odds_train)-1000, len(df_odds_train))\n",
        "\n",
        "\n",
        "# Final chosen/engineered features\n",
        "\n",
        "feats_odds_cat = ['event_team', 'opponent', 'league', 'side']\n",
        "d_cat = dict.fromkeys(feats_odds_cat, 'category')\n",
        "df_odds_train = df_odds_train.astype(d_cat)\n",
        "df_odds_train.info()\n",
        "\n",
        "feats_odds_num =['odd_h', 'odd_d', 'odd_a', 'diff_h_d', 'diff_d_a', 'diff_h_a']\n",
        "\n",
        "target_var_odds = 'result'\n",
        "\n",
        "procs = [FillMissing, Categorify, Normalize]\n",
        "\n",
        "test_odds = TabularList.from_df(df_odds_train.iloc[valid_idx_odds].copy(), cat_names = feats_odds_cat, cont_names= feats_odds_num)\n",
        "\n",
        "data_odds = (TabularList.from_df(df_odds_train, cat_names = feats_odds_cat, cont_names= feats_odds_num)\n",
        "                   .split_by_idx(valid_idx_odds)\n",
        "                   .label_from_df(cols=target_var_odds)\n",
        "                   .add_test(test_odds)        \n",
        "                   .databunch())\n",
        "\n",
        "\n",
        "\n",
        "# data = TabularDataBunch.from_df(path, df_train, target_var, valid_idx=valid_idx, cat_names=feats_cat)\n",
        "print(data_odds.train_ds.cont_names)  # `cont_names` defaults to: set(df)-set(cat_names)-{dep_var}"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 8089 entries, 8556 to 2732\n",
            "Data columns (total 11 columns):\n",
            "odd_h         8089 non-null float64\n",
            "odd_d         8089 non-null float64\n",
            "odd_a         8089 non-null float64\n",
            "event_team    8089 non-null category\n",
            "opponent      8089 non-null category\n",
            "league        8089 non-null category\n",
            "side          8089 non-null category\n",
            "diff_h_d      8089 non-null float64\n",
            "diff_d_a      8089 non-null float64\n",
            "diff_h_a      8089 non-null float64\n",
            "result        8089 non-null category\n",
            "dtypes: category(5), float64(6)\n",
            "memory usage: 488.2 KB\n",
            "['odd_h', 'odd_d', 'odd_a', 'diff_h_d', 'diff_d_a', 'diff_h_a']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "nLlaAnezURKM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 2302
        },
        "outputId": "b77af062-7cc1-447d-efc7-10c6847ae979"
      },
      "cell_type": "code",
      "source": [
        "learn = tabular_learner(data_odds, layers=[200, 200, 100], metrics=accuracy)\n",
        "# find best lr\n",
        "learn.lr_find()\n",
        "\n",
        "# find best lr\n",
        "learn.recorder.plot()\n",
        "\n",
        "# Train the network\n",
        "learn.fit(100, 1e-1)"
      ],
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "    <div>\n",
              "        <style>\n",
              "            /* Turns off some styling */\n",
              "            progress {\n",
              "                /* gets rid of default border in Firefox and Opera. */\n",
              "                border: none;\n",
              "                /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "                background-size: auto;\n",
              "            }\n",
              "            .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "                background: #F44336;\n",
              "            }\n",
              "        </style>\n",
              "      <progress value='0' class='' max='1', style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
              "      0.00% [0/1 00:00<00:00]\n",
              "    </div>\n",
              "    \n",
              "<table style='width:300px; margin-bottom:10px'>\n",
              "  <tr>\n",
              "    <th>epoch</th>\n",
              "    <th>train_loss</th>\n",
              "    <th>valid_loss</th>\n",
              "    <th>accuracy</th>\n",
              "  </tr>\n",
              "</table>\n",
              "\n",
              "\n",
              "    <div>\n",
              "        <style>\n",
              "            /* Turns off some styling */\n",
              "            progress {\n",
              "                /* gets rid of default border in Firefox and Opera. */\n",
              "                border: none;\n",
              "                /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "                background-size: auto;\n",
              "            }\n",
              "            .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "                background: #F44336;\n",
              "            }\n",
              "        </style>\n",
              "      <progress value='0' class='progress-bar-interrupted' max='111', style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
              "      Interrupted\n",
              "    </div>\n",
              "    "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "Total time: 04:11 <p><table style='width:300px; margin-bottom:10px'>\n",
              "  <tr>\n",
              "    <th>epoch</th>\n",
              "    <th>train_loss</th>\n",
              "    <th>valid_loss</th>\n",
              "    <th>accuracy</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>1</th>\n",
              "    <th>1.033755</th>\n",
              "    <th>0.986478</th>\n",
              "    <th>0.529000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>2</th>\n",
              "    <th>1.015018</th>\n",
              "    <th>9.891528</th>\n",
              "    <th>0.514000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>3</th>\n",
              "    <th>1.020972</th>\n",
              "    <th>11.066001</th>\n",
              "    <th>0.445000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>4</th>\n",
              "    <th>1.015994</th>\n",
              "    <th>9.405920</th>\n",
              "    <th>0.518000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>5</th>\n",
              "    <th>1.005765</th>\n",
              "    <th>0.988190</th>\n",
              "    <th>0.516000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>6</th>\n",
              "    <th>1.001960</th>\n",
              "    <th>24.589720</th>\n",
              "    <th>0.476000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>7</th>\n",
              "    <th>1.010902</th>\n",
              "    <th>0.989184</th>\n",
              "    <th>0.503000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>8</th>\n",
              "    <th>1.010024</th>\n",
              "    <th>3.600037</th>\n",
              "    <th>0.492000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>9</th>\n",
              "    <th>1.006931</th>\n",
              "    <th>8.796710</th>\n",
              "    <th>0.518000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>10</th>\n",
              "    <th>1.015065</th>\n",
              "    <th>1.022232</th>\n",
              "    <th>0.505000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>11</th>\n",
              "    <th>1.003981</th>\n",
              "    <th>5.931263</th>\n",
              "    <th>0.508000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>12</th>\n",
              "    <th>1.022524</th>\n",
              "    <th>1.594972</th>\n",
              "    <th>0.308000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>13</th>\n",
              "    <th>1.010237</th>\n",
              "    <th>17.322397</th>\n",
              "    <th>0.456000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>14</th>\n",
              "    <th>1.044311</th>\n",
              "    <th>17631.029297</th>\n",
              "    <th>0.515000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>15</th>\n",
              "    <th>1.011127</th>\n",
              "    <th>1.388166</th>\n",
              "    <th>0.506000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>16</th>\n",
              "    <th>1.008924</th>\n",
              "    <th>78890.460938</th>\n",
              "    <th>0.502000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>17</th>\n",
              "    <th>0.996090</th>\n",
              "    <th>91224.976562</th>\n",
              "    <th>0.442000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>18</th>\n",
              "    <th>1.007864</th>\n",
              "    <th>1.301234</th>\n",
              "    <th>0.516000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>19</th>\n",
              "    <th>1.010140</th>\n",
              "    <th>47.679165</th>\n",
              "    <th>0.488000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>20</th>\n",
              "    <th>1.012034</th>\n",
              "    <th>0.967351</th>\n",
              "    <th>0.524000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>21</th>\n",
              "    <th>1.002441</th>\n",
              "    <th>1.142712</th>\n",
              "    <th>0.521000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>22</th>\n",
              "    <th>0.994467</th>\n",
              "    <th>25.065035</th>\n",
              "    <th>0.503000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>23</th>\n",
              "    <th>1.007448</th>\n",
              "    <th>1222.343262</th>\n",
              "    <th>0.432000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>24</th>\n",
              "    <th>1.006013</th>\n",
              "    <th>841.350586</th>\n",
              "    <th>0.521000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>25</th>\n",
              "    <th>1.008777</th>\n",
              "    <th>444.820251</th>\n",
              "    <th>0.523000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>26</th>\n",
              "    <th>0.998039</th>\n",
              "    <th>0.975982</th>\n",
              "    <th>0.526000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>27</th>\n",
              "    <th>0.990452</th>\n",
              "    <th>1.097215</th>\n",
              "    <th>0.486000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>28</th>\n",
              "    <th>0.989965</th>\n",
              "    <th>1.048532</th>\n",
              "    <th>0.510000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>29</th>\n",
              "    <th>1.013203</th>\n",
              "    <th>19047.148438</th>\n",
              "    <th>0.499000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>30</th>\n",
              "    <th>0.995809</th>\n",
              "    <th>1145.710693</th>\n",
              "    <th>0.521000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>31</th>\n",
              "    <th>1.000644</th>\n",
              "    <th>1.083015</th>\n",
              "    <th>0.431000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>32</th>\n",
              "    <th>1.000336</th>\n",
              "    <th>201.875275</th>\n",
              "    <th>0.466000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>33</th>\n",
              "    <th>0.994036</th>\n",
              "    <th>355.611572</th>\n",
              "    <th>0.385000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>34</th>\n",
              "    <th>0.985061</th>\n",
              "    <th>1.181107</th>\n",
              "    <th>0.377000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>35</th>\n",
              "    <th>0.986305</th>\n",
              "    <th>1.063873</th>\n",
              "    <th>0.510000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>36</th>\n",
              "    <th>1.005467</th>\n",
              "    <th>24.581242</th>\n",
              "    <th>0.517000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>37</th>\n",
              "    <th>0.983969</th>\n",
              "    <th>1.041132</th>\n",
              "    <th>0.515000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>38</th>\n",
              "    <th>0.991540</th>\n",
              "    <th>5052.191406</th>\n",
              "    <th>0.510000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>39</th>\n",
              "    <th>0.983532</th>\n",
              "    <th>1.924646</th>\n",
              "    <th>0.511000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>40</th>\n",
              "    <th>1.014957</th>\n",
              "    <th>6496.436035</th>\n",
              "    <th>0.477000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>41</th>\n",
              "    <th>0.983183</th>\n",
              "    <th>1.016195</th>\n",
              "    <th>0.480000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>42</th>\n",
              "    <th>0.999580</th>\n",
              "    <th>3378.064697</th>\n",
              "    <th>0.520000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>43</th>\n",
              "    <th>1.007721</th>\n",
              "    <th>13717.788086</th>\n",
              "    <th>0.516000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>44</th>\n",
              "    <th>0.988535</th>\n",
              "    <th>0.978413</th>\n",
              "    <th>0.514000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>45</th>\n",
              "    <th>0.982934</th>\n",
              "    <th>0.993627</th>\n",
              "    <th>0.522000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>46</th>\n",
              "    <th>0.990668</th>\n",
              "    <th>1.041322</th>\n",
              "    <th>0.520000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>47</th>\n",
              "    <th>0.975943</th>\n",
              "    <th>18.354422</th>\n",
              "    <th>0.525000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>48</th>\n",
              "    <th>1.000002</th>\n",
              "    <th>56.782440</th>\n",
              "    <th>0.456000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>49</th>\n",
              "    <th>0.993507</th>\n",
              "    <th>84.739449</th>\n",
              "    <th>0.517000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>50</th>\n",
              "    <th>0.997068</th>\n",
              "    <th>2216.222656</th>\n",
              "    <th>0.493000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>51</th>\n",
              "    <th>0.990918</th>\n",
              "    <th>1.006870</th>\n",
              "    <th>0.514000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>52</th>\n",
              "    <th>0.988593</th>\n",
              "    <th>0.984018</th>\n",
              "    <th>0.503000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>53</th>\n",
              "    <th>0.996263</th>\n",
              "    <th>2.332873</th>\n",
              "    <th>0.453000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>54</th>\n",
              "    <th>1.010349</th>\n",
              "    <th>1.661602</th>\n",
              "    <th>0.464000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>55</th>\n",
              "    <th>0.979618</th>\n",
              "    <th>1.017624</th>\n",
              "    <th>0.521000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>56</th>\n",
              "    <th>0.998257</th>\n",
              "    <th>1122.872070</th>\n",
              "    <th>0.514000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>57</th>\n",
              "    <th>0.982225</th>\n",
              "    <th>8508.758789</th>\n",
              "    <th>0.469000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>58</th>\n",
              "    <th>0.986777</th>\n",
              "    <th>4621.997559</th>\n",
              "    <th>0.487000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>59</th>\n",
              "    <th>0.982592</th>\n",
              "    <th>1.078472</th>\n",
              "    <th>0.429000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>60</th>\n",
              "    <th>0.989926</th>\n",
              "    <th>50.281624</th>\n",
              "    <th>0.427000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>61</th>\n",
              "    <th>0.985196</th>\n",
              "    <th>8620.523438</th>\n",
              "    <th>0.508000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>62</th>\n",
              "    <th>0.982547</th>\n",
              "    <th>678.632568</th>\n",
              "    <th>0.448000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>63</th>\n",
              "    <th>0.982574</th>\n",
              "    <th>1.375281</th>\n",
              "    <th>0.473000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>64</th>\n",
              "    <th>0.988809</th>\n",
              "    <th>0.991113</th>\n",
              "    <th>0.512000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>65</th>\n",
              "    <th>1.012369</th>\n",
              "    <th>983.607361</th>\n",
              "    <th>0.506000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>66</th>\n",
              "    <th>0.985043</th>\n",
              "    <th>1.923675</th>\n",
              "    <th>0.501000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>67</th>\n",
              "    <th>0.976502</th>\n",
              "    <th>145.733536</th>\n",
              "    <th>0.502000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>68</th>\n",
              "    <th>0.987959</th>\n",
              "    <th>1.045267</th>\n",
              "    <th>0.451000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>69</th>\n",
              "    <th>0.984613</th>\n",
              "    <th>12.359727</th>\n",
              "    <th>0.501000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>70</th>\n",
              "    <th>1.003112</th>\n",
              "    <th>1.268126</th>\n",
              "    <th>0.509000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>71</th>\n",
              "    <th>0.988377</th>\n",
              "    <th>4107.168945</th>\n",
              "    <th>0.495000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>72</th>\n",
              "    <th>0.981072</th>\n",
              "    <th>71942.789062</th>\n",
              "    <th>0.507000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>73</th>\n",
              "    <th>0.985750</th>\n",
              "    <th>47572.613281</th>\n",
              "    <th>0.489000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>74</th>\n",
              "    <th>0.975124</th>\n",
              "    <th>96334.734375</th>\n",
              "    <th>0.482000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>75</th>\n",
              "    <th>0.978196</th>\n",
              "    <th>353.825531</th>\n",
              "    <th>0.516000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>76</th>\n",
              "    <th>0.988221</th>\n",
              "    <th>218.358734</th>\n",
              "    <th>0.498000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>77</th>\n",
              "    <th>0.983819</th>\n",
              "    <th>90376.625000</th>\n",
              "    <th>0.497000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>78</th>\n",
              "    <th>0.982969</th>\n",
              "    <th>7015.290527</th>\n",
              "    <th>0.522000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>79</th>\n",
              "    <th>0.982483</th>\n",
              "    <th>526.484741</th>\n",
              "    <th>0.511000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>80</th>\n",
              "    <th>0.995655</th>\n",
              "    <th>18.398924</th>\n",
              "    <th>0.415000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>81</th>\n",
              "    <th>0.980884</th>\n",
              "    <th>13668.501953</th>\n",
              "    <th>0.505000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>82</th>\n",
              "    <th>1.000481</th>\n",
              "    <th>244.477936</th>\n",
              "    <th>0.470000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>83</th>\n",
              "    <th>0.991745</th>\n",
              "    <th>1.113451</th>\n",
              "    <th>0.519000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>84</th>\n",
              "    <th>1.007313</th>\n",
              "    <th>307.537903</th>\n",
              "    <th>0.450000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>85</th>\n",
              "    <th>0.991023</th>\n",
              "    <th>1638.682617</th>\n",
              "    <th>0.516000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>86</th>\n",
              "    <th>0.969282</th>\n",
              "    <th>1.919319</th>\n",
              "    <th>0.521000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>87</th>\n",
              "    <th>0.979339</th>\n",
              "    <th>530.138672</th>\n",
              "    <th>0.512000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>88</th>\n",
              "    <th>0.977483</th>\n",
              "    <th>595.470337</th>\n",
              "    <th>0.523000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>89</th>\n",
              "    <th>0.995975</th>\n",
              "    <th>47.647392</th>\n",
              "    <th>0.520000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>90</th>\n",
              "    <th>0.998967</th>\n",
              "    <th>1.091742</th>\n",
              "    <th>0.467000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>91</th>\n",
              "    <th>0.981450</th>\n",
              "    <th>0.989272</th>\n",
              "    <th>0.493000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>92</th>\n",
              "    <th>0.994561</th>\n",
              "    <th>8.514512</th>\n",
              "    <th>0.495000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>93</th>\n",
              "    <th>0.980019</th>\n",
              "    <th>8.237616</th>\n",
              "    <th>0.447000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>94</th>\n",
              "    <th>1.003452</th>\n",
              "    <th>3918.088379</th>\n",
              "    <th>0.511000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>95</th>\n",
              "    <th>0.996893</th>\n",
              "    <th>5.938885</th>\n",
              "    <th>0.349000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>96</th>\n",
              "    <th>0.980062</th>\n",
              "    <th>169.377213</th>\n",
              "    <th>0.510000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>97</th>\n",
              "    <th>0.981187</th>\n",
              "    <th>405.760223</th>\n",
              "    <th>0.507000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>98</th>\n",
              "    <th>1.005961</th>\n",
              "    <th>1.562848</th>\n",
              "    <th>0.336000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>99</th>\n",
              "    <th>0.995490</th>\n",
              "    <th>15.753776</th>\n",
              "    <th>0.480000</th>\n",
              "  </tr>\n",
              "  <tr>\n",
              "    <th>100</th>\n",
              "    <th>0.975338</th>\n",
              "    <th>3.550497</th>\n",
              "    <th>0.522000</th>\n",
              "  </tr>\n",
              "</table>\n"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "Qi5K9iE4Uy-f",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}