videogames_full_GlobalSales.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "videogames_full_GlobalSales.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyPeBBdfQ4qdz+P2UyvQa5LQ",
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/johanv26/Analitica/blob/main/videogames_full_GlobalSales.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kSfo5Wp8d-kv",
        "outputId": "efd87950-705f-4b0d-c895-e504c95b3f0f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "#descargar datos video_games\n",
        "!pip install wget\n",
        "import wget\n",
        "import pandas as pd\n",
        "#datos video juelgos-2016\n",
        "#csv file id#\n",
        "FILE_ID='1kMvhv1VpUpGmyyDZneZ7DaIpXZx9826Z'\n",
        "url='https://drive.google.com/uc?export=download&id='+FILE_ID\n",
        "#datos Colombia\n",
        "import os.path as path\n",
        "csv_file='Video_Games_Sales_as_at_22_Dec_2016.csv'\n",
        "if not path.exists(csv_file):\n",
        "  wget.download(url,csv_file)\n",
        "\n",
        "gamesdf = pd.read_csv(csv_file)\n",
        "gamesdf.head(20)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: wget in /usr/local/lib/python3.6/dist-packages (3.2)\n"
          ],
          "name": "stdout"
        },
        {
          "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>Name</th>\n",
              "      <th>Platform</th>\n",
              "      <th>Year_of_Release</th>\n",
              "      <th>Genre</th>\n",
              "      <th>Publisher</th>\n",
              "      <th>NA_Sales</th>\n",
              "      <th>EU_Sales</th>\n",
              "      <th>JP_Sales</th>\n",
              "      <th>Other_Sales</th>\n",
              "      <th>Global_Sales</th>\n",
              "      <th>Critic_Score</th>\n",
              "      <th>Critic_Count</th>\n",
              "      <th>User_Score</th>\n",
              "      <th>User_Count</th>\n",
              "      <th>Developer</th>\n",
              "      <th>Rating</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Wii Sports</td>\n",
              "      <td>Wii</td>\n",
              "      <td>2006.0</td>\n",
              "      <td>Sports</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>41.36</td>\n",
              "      <td>28.96</td>\n",
              "      <td>3.77</td>\n",
              "      <td>8.45</td>\n",
              "      <td>82.53</td>\n",
              "      <td>76.0</td>\n",
              "      <td>51.0</td>\n",
              "      <td>8</td>\n",
              "      <td>322.0</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Super Mario Bros.</td>\n",
              "      <td>NES</td>\n",
              "      <td>1985.0</td>\n",
              "      <td>Platform</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>29.08</td>\n",
              "      <td>3.58</td>\n",
              "      <td>6.81</td>\n",
              "      <td>0.77</td>\n",
              "      <td>40.24</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Mario Kart Wii</td>\n",
              "      <td>Wii</td>\n",
              "      <td>2008.0</td>\n",
              "      <td>Racing</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>15.68</td>\n",
              "      <td>12.76</td>\n",
              "      <td>3.79</td>\n",
              "      <td>3.29</td>\n",
              "      <td>35.52</td>\n",
              "      <td>82.0</td>\n",
              "      <td>73.0</td>\n",
              "      <td>8.3</td>\n",
              "      <td>709.0</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Wii Sports Resort</td>\n",
              "      <td>Wii</td>\n",
              "      <td>2009.0</td>\n",
              "      <td>Sports</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>15.61</td>\n",
              "      <td>10.93</td>\n",
              "      <td>3.28</td>\n",
              "      <td>2.95</td>\n",
              "      <td>32.77</td>\n",
              "      <td>80.0</td>\n",
              "      <td>73.0</td>\n",
              "      <td>8</td>\n",
              "      <td>192.0</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Pokemon Red/Pokemon Blue</td>\n",
              "      <td>GB</td>\n",
              "      <td>1996.0</td>\n",
              "      <td>Role-Playing</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>11.27</td>\n",
              "      <td>8.89</td>\n",
              "      <td>10.22</td>\n",
              "      <td>1.00</td>\n",
              "      <td>31.37</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>Tetris</td>\n",
              "      <td>GB</td>\n",
              "      <td>1989.0</td>\n",
              "      <td>Puzzle</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>23.20</td>\n",
              "      <td>2.26</td>\n",
              "      <td>4.22</td>\n",
              "      <td>0.58</td>\n",
              "      <td>30.26</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>New Super Mario Bros.</td>\n",
              "      <td>DS</td>\n",
              "      <td>2006.0</td>\n",
              "      <td>Platform</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>11.28</td>\n",
              "      <td>9.14</td>\n",
              "      <td>6.50</td>\n",
              "      <td>2.88</td>\n",
              "      <td>29.80</td>\n",
              "      <td>89.0</td>\n",
              "      <td>65.0</td>\n",
              "      <td>8.5</td>\n",
              "      <td>431.0</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>Wii Play</td>\n",
              "      <td>Wii</td>\n",
              "      <td>2006.0</td>\n",
              "      <td>Misc</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>13.96</td>\n",
              "      <td>9.18</td>\n",
              "      <td>2.93</td>\n",
              "      <td>2.84</td>\n",
              "      <td>28.92</td>\n",
              "      <td>58.0</td>\n",
              "      <td>41.0</td>\n",
              "      <td>6.6</td>\n",
              "      <td>129.0</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>New Super Mario Bros. Wii</td>\n",
              "      <td>Wii</td>\n",
              "      <td>2009.0</td>\n",
              "      <td>Platform</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>14.44</td>\n",
              "      <td>6.94</td>\n",
              "      <td>4.70</td>\n",
              "      <td>2.24</td>\n",
              "      <td>28.32</td>\n",
              "      <td>87.0</td>\n",
              "      <td>80.0</td>\n",
              "      <td>8.4</td>\n",
              "      <td>594.0</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>Duck Hunt</td>\n",
              "      <td>NES</td>\n",
              "      <td>1984.0</td>\n",
              "      <td>Shooter</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>26.93</td>\n",
              "      <td>0.63</td>\n",
              "      <td>0.28</td>\n",
              "      <td>0.47</td>\n",
              "      <td>28.31</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>Nintendogs</td>\n",
              "      <td>DS</td>\n",
              "      <td>2005.0</td>\n",
              "      <td>Simulation</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>9.05</td>\n",
              "      <td>10.95</td>\n",
              "      <td>1.93</td>\n",
              "      <td>2.74</td>\n",
              "      <td>24.67</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>Mario Kart DS</td>\n",
              "      <td>DS</td>\n",
              "      <td>2005.0</td>\n",
              "      <td>Racing</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>9.71</td>\n",
              "      <td>7.47</td>\n",
              "      <td>4.13</td>\n",
              "      <td>1.90</td>\n",
              "      <td>23.21</td>\n",
              "      <td>91.0</td>\n",
              "      <td>64.0</td>\n",
              "      <td>8.6</td>\n",
              "      <td>464.0</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>Pokemon Gold/Pokemon Silver</td>\n",
              "      <td>GB</td>\n",
              "      <td>1999.0</td>\n",
              "      <td>Role-Playing</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>9.00</td>\n",
              "      <td>6.18</td>\n",
              "      <td>7.20</td>\n",
              "      <td>0.71</td>\n",
              "      <td>23.10</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>Wii Fit</td>\n",
              "      <td>Wii</td>\n",
              "      <td>2007.0</td>\n",
              "      <td>Sports</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>8.92</td>\n",
              "      <td>8.03</td>\n",
              "      <td>3.60</td>\n",
              "      <td>2.15</td>\n",
              "      <td>22.70</td>\n",
              "      <td>80.0</td>\n",
              "      <td>63.0</td>\n",
              "      <td>7.7</td>\n",
              "      <td>146.0</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>Kinect Adventures!</td>\n",
              "      <td>X360</td>\n",
              "      <td>2010.0</td>\n",
              "      <td>Misc</td>\n",
              "      <td>Microsoft Game Studios</td>\n",
              "      <td>15.00</td>\n",
              "      <td>4.89</td>\n",
              "      <td>0.24</td>\n",
              "      <td>1.69</td>\n",
              "      <td>21.81</td>\n",
              "      <td>61.0</td>\n",
              "      <td>45.0</td>\n",
              "      <td>6.3</td>\n",
              "      <td>106.0</td>\n",
              "      <td>Good Science Studio</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>Wii Fit Plus</td>\n",
              "      <td>Wii</td>\n",
              "      <td>2009.0</td>\n",
              "      <td>Sports</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>9.01</td>\n",
              "      <td>8.49</td>\n",
              "      <td>2.53</td>\n",
              "      <td>1.77</td>\n",
              "      <td>21.79</td>\n",
              "      <td>80.0</td>\n",
              "      <td>33.0</td>\n",
              "      <td>7.4</td>\n",
              "      <td>52.0</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>Grand Theft Auto V</td>\n",
              "      <td>PS3</td>\n",
              "      <td>2013.0</td>\n",
              "      <td>Action</td>\n",
              "      <td>Take-Two Interactive</td>\n",
              "      <td>7.02</td>\n",
              "      <td>9.09</td>\n",
              "      <td>0.98</td>\n",
              "      <td>3.96</td>\n",
              "      <td>21.04</td>\n",
              "      <td>97.0</td>\n",
              "      <td>50.0</td>\n",
              "      <td>8.2</td>\n",
              "      <td>3994.0</td>\n",
              "      <td>Rockstar North</td>\n",
              "      <td>M</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>Grand Theft Auto: San Andreas</td>\n",
              "      <td>PS2</td>\n",
              "      <td>2004.0</td>\n",
              "      <td>Action</td>\n",
              "      <td>Take-Two Interactive</td>\n",
              "      <td>9.43</td>\n",
              "      <td>0.40</td>\n",
              "      <td>0.41</td>\n",
              "      <td>10.57</td>\n",
              "      <td>20.81</td>\n",
              "      <td>95.0</td>\n",
              "      <td>80.0</td>\n",
              "      <td>9</td>\n",
              "      <td>1588.0</td>\n",
              "      <td>Rockstar North</td>\n",
              "      <td>M</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>Super Mario World</td>\n",
              "      <td>SNES</td>\n",
              "      <td>1990.0</td>\n",
              "      <td>Platform</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>12.78</td>\n",
              "      <td>3.75</td>\n",
              "      <td>3.54</td>\n",
              "      <td>0.55</td>\n",
              "      <td>20.61</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>Brain Age: Train Your Brain in Minutes a Day</td>\n",
              "      <td>DS</td>\n",
              "      <td>2005.0</td>\n",
              "      <td>Misc</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>4.74</td>\n",
              "      <td>9.20</td>\n",
              "      <td>4.16</td>\n",
              "      <td>2.04</td>\n",
              "      <td>20.15</td>\n",
              "      <td>77.0</td>\n",
              "      <td>58.0</td>\n",
              "      <td>7.9</td>\n",
              "      <td>50.0</td>\n",
              "      <td>Nintendo</td>\n",
              "      <td>E</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                            Name  ... Rating\n",
              "0                                     Wii Sports  ...      E\n",
              "1                              Super Mario Bros.  ...    NaN\n",
              "2                                 Mario Kart Wii  ...      E\n",
              "3                              Wii Sports Resort  ...      E\n",
              "4                       Pokemon Red/Pokemon Blue  ...    NaN\n",
              "5                                         Tetris  ...    NaN\n",
              "6                          New Super Mario Bros.  ...      E\n",
              "7                                       Wii Play  ...      E\n",
              "8                      New Super Mario Bros. Wii  ...      E\n",
              "9                                      Duck Hunt  ...    NaN\n",
              "10                                    Nintendogs  ...    NaN\n",
              "11                                 Mario Kart DS  ...      E\n",
              "12                   Pokemon Gold/Pokemon Silver  ...    NaN\n",
              "13                                       Wii Fit  ...      E\n",
              "14                            Kinect Adventures!  ...      E\n",
              "15                                  Wii Fit Plus  ...      E\n",
              "16                            Grand Theft Auto V  ...      M\n",
              "17                 Grand Theft Auto: San Andreas  ...      M\n",
              "18                             Super Mario World  ...    NaN\n",
              "19  Brain Age: Train Your Brain in Minutes a Day  ...      E\n",
              "\n",
              "[20 rows x 16 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NWK4HczeeMBy",
        "outputId": "fb677381-7d74-4401-d14c-fa1c884a55f3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 230
        }
      },
      "source": [
        "#Paso 2: Particion entrenamiento y validacion\n",
        "#Tamanio Xtrain 70%. Tamanio Xtest=30%\n",
        "from sklearn.model_selection import train_test_split\n",
        "import warnings\n",
        "from sklearn.impute import SimpleImputer \n",
        "from sklearn.model_selection import train_test_split, GridSearchCV,cross_val_score, cross_val_predict\n",
        "from sklearn.metrics import mean_absolute_error\n",
        "from sklearn.pipeline import Pipeline\n",
        "\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "df_train, df_test =train_test_split(gamesdf,test_size=0.3)\n",
        "col_sal= 'Year_of_Release'\n",
        "ytrain=df_train[col_sal]\n",
        "ytest=df_test[col_sal]\n",
        "df_train.drop(columns=col_sal,inplace=True)\n",
        "df_test.drop(columns=col_sal,inplace=True)\n",
        "ytrain"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "5009     1996.0\n",
              "11339       NaN\n",
              "11143    2011.0\n",
              "16427    2016.0\n",
              "176      2012.0\n",
              "          ...  \n",
              "6089     1997.0\n",
              "6808     2008.0\n",
              "10037    2004.0\n",
              "11894    2011.0\n",
              "2247     2015.0\n",
              "Name: Year_of_Release, Length: 11703, dtype: float64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qlroDgFveO9A",
        "outputId": "939de7e2-6676-4ffa-fcd7-eb705a934dde",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "#Completao de datos por moda de la salida Global sales\n",
        "from sklearn.impute import SimpleImputer\n",
        "imputer=SimpleImputer(strategy='most_frequent')\n",
        "ytrain=imputer.fit_transform(pd.DataFrame(ytrain))\n",
        "ytrain=ytrain.reshape(-1)\n",
        "\n",
        "ytest=imputer.transform(pd.DataFrame(ytest))\n",
        "ytest=ytest.reshape(-1)\n",
        "\n",
        "print(ytrain.shape,ytest.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(11703,) (5016,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XtPatvHreQ4a"
      },
      "source": [
        "import numpy as np"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "taxDCY2ceRxw",
        "outputId": "ced32ab9-140f-4058-ca4c-5ea39f4a0c2a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 513
        }
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.hist(ytrain,bins=100)\n",
        "plt.show()\n",
        "\n",
        "#diagrama de caja\n",
        "\n",
        "#caja->\n",
        "plt.boxplot(ytrain)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "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"
          }
        },
        {
          "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": "Di3e2GkqeYDF"
      },
      "source": [
        "#definir columnas \n",
        "col_scores = ['User_Score','Critic_Score','Critic_Count','User_Count']\n",
        "# Strange values tbd\n",
        "\n",
        "cat = ['Name','Platform','Genre','Publisher','Developer','Rating']\n",
        "items = [] \n",
        "for i in cat:\n",
        "    items += [list(gamesdf[i].value_counts().index)]\n",
        "cat_names = dict(zip(cat, items))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ud43r7HzeekW"
      },
      "source": [
        "\n",
        "\n",
        "```\n",
        "# Tiene formato de código\n",
        "```\n",
        "\n",
        " #Crear clase propia de preproceso para video juegos"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zJaVhGEoebx9"
      },
      "source": [
        "#%% dummy transformer\n",
        "from sklearn.base import BaseEstimator\n",
        "from sklearn.base import TransformerMixin\n",
        "from sklearn.preprocessing import OrdinalEncoder,OneHotEncoder\n",
        "class dummy_games(BaseEstimator,TransformerMixin):\n",
        "    #inicializacion de clase y varaibles\n",
        "    def __init__(self,col_scores, cat_usr):#constructor clase\n",
        "        \n",
        "        self.col_stats = col_scores#lista atributos tipo score\n",
        "        self.cat_usr = cat_usr #lista de atributos categoricos\n",
        "\n",
        "    def fit(self,X, *_):\n",
        "        Xi = X.copy() #copiar dataset para no reemplazar original\n",
        "        self.imputer_num = SimpleImputer(strategy=\"most_frequent\")#crear imputador tipo modo\n",
        "        self.a = Xi.columns[np.sum(Xi.isna())> 0] #encontrar columnas con datos faltantes\n",
        "        #print(a)\n",
        "        self.imputer_num.fit(Xi[self.a])# ajustar imputador\n",
        "        \n",
        "        Xi[self.a] = self.imputer_num.transform(Xi[self.a]) #evaluar  sobre datos imputador\n",
        "        \n",
        "        #codificar\n",
        "        #only User_Score is cnahge the tbd most frequetn for mean.\n",
        "        frec_val=gamesdf['User_Score'].value_counts()\n",
        "        mean=np.mean(frec_val.index[1:].astype('float64')) \n",
        "        Xi['User_Score'].replace(regex=[\"tbd\"], value=mean,inplace=True)      \n",
        "        Xi['User_Score'] = Xi['User_Score'].astype('float64');\n",
        "        \n",
        "\n",
        "        \n",
        "        \n",
        "        \n",
        "        cat = []#codificar variables categoricas con ordinal encoder\n",
        "        for i in self.cat_usr.keys():\n",
        "            cat = cat + [[*self.cat_usr.get(i)]]\n",
        "        self.col_cat_usr = OrdinalEncoder(categories=cat)\n",
        "        Xi[[*self.cat_usr.keys()]] =self.col_cat_usr.fit_transform(Xi[[*self.cat_usr.keys()]])\n",
        "        \n",
        "\n",
        "        \n",
        "        return self    \n",
        "\n",
        "    def transform(self, X, *_):\n",
        "        Xi = X.copy()\n",
        "        Xi[self.a] = self.imputer_num.transform(Xi[self.a])\n",
        "\n",
        "              \n",
        "        frec_val=gamesdf['User_Score'].value_counts()\n",
        "        mean=np.mean(frec_val.index[1:].astype('float64')) \n",
        "        Xi['User_Score'].replace(regex=[\"tbd\"], value=mean,inplace=True)      \n",
        "        Xi['User_Score'] = Xi['User_Score'].astype('float64');\n",
        "          \n",
        "        Xi[[*self.cat_usr.keys()]] =self.col_cat_usr.transform(Xi[[*self.cat_usr.keys()]])\n",
        "        \n",
        "        #Fase de transformacion\n",
        "        #\n",
        "        \n",
        "        return Xi\n",
        "    \n",
        "    def fit_transform(self,X,*_):\n",
        "        self.fit(X)\n",
        "        return self.transform(X)\n",
        "#%%"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "--xHifVeekqE",
        "outputId": "af627ed0-ee53-4768-b287-e7fbaeed4139",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 408
        }
      },
      "source": [
        "dummy=dummy_games(col_scores=col_scores,cat_usr=cat_names)\n",
        "dftrain_pre=dummy.fit_transform(df_train)\n",
        "dftrain_pre.info()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 11703 entries, 5009 to 2247\n",
            "Data columns (total 15 columns):\n",
            " #   Column        Non-Null Count  Dtype  \n",
            "---  ------        --------------  -----  \n",
            " 0   Name          11703 non-null  float64\n",
            " 1   Platform      11703 non-null  float64\n",
            " 2   Genre         11703 non-null  float64\n",
            " 3   Publisher     11703 non-null  float64\n",
            " 4   NA_Sales      11703 non-null  float64\n",
            " 5   EU_Sales      11703 non-null  float64\n",
            " 6   JP_Sales      11703 non-null  float64\n",
            " 7   Other_Sales   11703 non-null  float64\n",
            " 8   Global_Sales  11703 non-null  float64\n",
            " 9   Critic_Score  11703 non-null  float64\n",
            " 10  Critic_Count  11703 non-null  float64\n",
            " 11  User_Score    11703 non-null  float64\n",
            " 12  User_Count    11703 non-null  float64\n",
            " 13  Developer     11703 non-null  float64\n",
            " 14  Rating        11703 non-null  float64\n",
            "dtypes: float64(15)\n",
            "memory usage: 1.4 MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CQ14Tjl4emwA"
      },
      "source": [
        "# Predicción variable de interés desde datos preprocesados"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cE61YSZlemcJ"
      },
      "source": [
        "from sklearn.decomposition import PCA \n",
        "from sklearn.preprocessing import StandardScaler\n",
        "sca = StandardScaler()\n",
        "dftrain_pre_z = sca.fit_transform(dftrain_pre)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CdaXawGqfIcB",
        "outputId": "8b2ab74c-ce3b-4106-ba85-7a8680d22f86",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 248
        }
      },
      "source": [
        "dftrain_pre_z"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[-0.40560691,  0.04469935, -1.1516564 , ..., -0.18357   ,\n",
              "        -0.54650801, -0.63293221],\n",
              "       [-0.35593966, -0.92439375,  1.24818955, ..., -0.18357   ,\n",
              "        -0.52828737, -0.63293221],\n",
              "       [-0.9022794 , -0.53675651, -0.55169491, ..., -0.18357   ,\n",
              "        -0.52828737,  2.46977889],\n",
              "       ...,\n",
              "       [-0.52531056,  0.43233659,  0.04826658, ..., -0.18357   ,\n",
              "         3.6533485 ,  0.40130483],\n",
              "       [-0.88414249, -0.34293789,  0.64822807, ..., -0.18357   ,\n",
              "        -0.54650801, -0.63293221],\n",
              "       [-0.92599691,  1.40142969, -0.85167566, ...,  0.28249568,\n",
              "        -0.40074292, -0.63293221]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "An4Py-QXep9Y",
        "outputId": "fe5e7992-41a6-4c4b-f103-07f532ca3544",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "dftrain_pre_z.var(axis=0)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-SsVqNXKfrpZ",
        "outputId": "6f6fc475-932d-4082-8b3e-aba746f86c28",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        }
      },
      "source": [
        "dftrain_pre.var(axis=0)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Name            1.284508e+07\n",
              "Platform        2.662233e+01\n",
              "Genre           1.111349e+01\n",
              "Publisher       5.928002e+03\n",
              "NA_Sales        6.840003e-01\n",
              "EU_Sales        2.582391e-01\n",
              "JP_Sales        9.691093e-02\n",
              "Other_Sales     4.039379e-02\n",
              "Global_Sales    2.473329e+00\n",
              "Critic_Score    9.517153e+01\n",
              "Critic_Count    2.945463e+02\n",
              "User_Score      2.147158e+00\n",
              "User_Count      1.592828e+05\n",
              "Developer       1.084458e+05\n",
              "Rating          9.349684e-01\n",
              "dtype: float64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f2FXnCbLf4Dw"
      },
      "source": [
        "red = PCA()\n",
        "zz = red.fit_transform(dftrain_pre_z)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dj2LyPSAf5wY",
        "outputId": "ebbc0a4d-12ce-4412-d990-3a061f4ccc49",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "plt.scatter(zz[:,0],zz[:,1],c=ytrain,s = 100*dftrain_pre['EU_Sales']/(dftrain_pre['EU_Sales'].max()))\n",
        "plt.colorbar()\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0JInb6Amg6rP"
      },
      "source": [
        "# proceso con test\n",
        "dftest_pre = dummy.transform(df_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JILtNJvohWmJ",
        "outputId": "2d4d4551-0398-49cd-cea5-7569d5793b17",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 230
        }
      },
      "source": [
        "dftest_pre['EU_Sales']"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "12800    0.00\n",
              "15046    0.01\n",
              "9092     0.00\n",
              "2989     0.22\n",
              "6650     0.10\n",
              "         ... \n",
              "173      1.21\n",
              "6461     0.10\n",
              "5084     0.14\n",
              "11312    0.00\n",
              "16129    0.00\n",
              "Name: EU_Sales, Length: 5016, dtype: float64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nEZxxVwvhgAe",
        "outputId": "44c07988-d731-4b43-af76-7bc5e92ad4b8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 408
        }
      },
      "source": [
        "dftest_pre.info()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 5016 entries, 12800 to 16129\n",
            "Data columns (total 15 columns):\n",
            " #   Column        Non-Null Count  Dtype  \n",
            "---  ------        --------------  -----  \n",
            " 0   Name          5016 non-null   float64\n",
            " 1   Platform      5016 non-null   float64\n",
            " 2   Genre         5016 non-null   float64\n",
            " 3   Publisher     5016 non-null   float64\n",
            " 4   NA_Sales      5016 non-null   float64\n",
            " 5   EU_Sales      5016 non-null   float64\n",
            " 6   JP_Sales      5016 non-null   float64\n",
            " 7   Other_Sales   5016 non-null   float64\n",
            " 8   Global_Sales  5016 non-null   float64\n",
            " 9   Critic_Score  5016 non-null   float64\n",
            " 10  Critic_Count  5016 non-null   float64\n",
            " 11  User_Score    5016 non-null   float64\n",
            " 12  User_Count    5016 non-null   float64\n",
            " 13  Developer     5016 non-null   float64\n",
            " 14  Rating        5016 non-null   float64\n",
            "dtypes: float64(15)\n",
            "memory usage: 627.0 KB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6ZmlHn38h-7Q"
      },
      "source": [
        "#Xtest_pre_z = sca.transform(Xtest_pre) # standardizacion\n",
        "#zztest = red.transform(Xtest_pre_z) # llevarlo a 2D con pca\n",
        "\n",
        "zztest = red.transform(sca.transform(dummy.transform(df_test))) # una sola linea"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ad_LCp5ZiGx1",
        "outputId": "7dce952a-4324-4c34-b58c-aaea97303704",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "plt.scatter(zz[:,0],zz[:,1],c=ytrain,s = 100*dftrain_pre['EU_Sales']/(dftrain_pre['EU_Sales'].max()),label='train')\n",
        "plt.colorbar()\n",
        "plt.scatter(zztest[:,0],zztest[:,1],c=ytest,s=100*dftest_pre['EU_Sales']/dftest_pre['EU_Sales'].max(),marker='d',label='test')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DMnvpSMWiL0o"
      },
      "source": [
        "# definir modelos de predicción\n",
        "from sklearn.linear_model import LinearRegression, ElasticNet\n",
        "\n",
        "steps = [\n",
        "         [('scaler', StandardScaler()), #estandarizar cada atributo columna de xtrain centrada en 0 y var = 1\n",
        "          ('reg', LinearRegression())],\n",
        "         \n",
        "         [('scaler',StandardScaler()),\n",
        "          ('reg',ElasticNet())]\n",
        "         \n",
        "        # [('scaler', StandardScaler()), #estandarizar cada atriuto columna de xtrain centrada en 0 y var = 1\n",
        "        #  ('reg',  KernelRidge(kernel = 'rbf'))], #clasificador \n",
        "         ]        \n",
        "\n",
        "#parametros a buscar por busqueda por grilla\n",
        "parameters =[ \n",
        "             {'reg__fit_intercept':[True, False]             \n",
        "             },\n",
        "             {\n",
        "              'reg__alpha': [0,1e-3,1e-2,1e-1,1,10], #parametros n_neighbors debe ser siempre un int\n",
        "              'reg__l1_ratio':[0,0.25,0.5,0.75,1]\n",
        "             }\n",
        "              ]\n",
        "\n",
        "label_model = ['Nor+RegLin','Nor+ElasticNet']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nr48B0buiqf1",
        "outputId": "3d323b0c-fd85-4663-a037-4c86a395297b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 70
        }
      },
      "source": [
        "parameters"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'reg__fit_intercept': [True, False]},\n",
              " {'reg__alpha': [0, 0.001, 0.01, 0.1, 1, 10],\n",
              "  'reg__l1_ratio': [0, 0.25, 0.5, 0.75, 1]}]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ovMrX4sJixxs",
        "outputId": "4b1d31d3-31cc-4bbc-a24b-4f62a93ec484",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "import os\n",
        "pathpre = 'datospre'\n",
        "\n",
        "try:\n",
        "  os.mkdir(pathpre)\n",
        "except:\n",
        "  print(\"Carpeta results ya existe\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Carpeta results ya existe\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_GK56hvUjBqt",
        "outputId": "c8f06346-75e8-4af2-cfdf-b2ac5aa6c61d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "#valildacion cruzada anidada\n",
        "from joblib import dump, load\n",
        "from sklearn.metrics import mean_absolute_error as msa \n",
        "Niter = 10 #numero particiones outter loop nested cross-validation\n",
        "msev =np.zeros((Niter,len(steps)))#arreglo para guardar acierto/error\n",
        "Nmod = len(steps) #numero de modelos a probar\n",
        "best_estimators = Niter*[None]#mejor metodo por iteracion\n",
        "###clave del funcionamiento\n",
        "best_hyperpar = Niter*[None]#mejor metodo por iteracion \n",
        "\n",
        "\n",
        "#############################\n",
        "for j in range(Niter): #outter loop # SI TIENE MENOS DE 1000 DATOS BORRAR ESTE CICLO SOLO CV EN LINEA 16\n",
        "      #print('it %d/%d'%(j+1,Niter))\n",
        "      #particiono datos outter loop\n",
        "      X_trainj, X_testj, y_trainj, y_testj = train_test_split(dftrain_pre,ytrain,test_size=0.3) # xtrain 60, xtest 26\n",
        "      list_est = [] #lista lazo interno para guardar mejor estimador por modelo para iteracion j\n",
        "      list_hyper = [] #lista lazo interno para guardar mejores hyperparametros por modelo para iteracion j\n",
        "      for r in range(Nmod): #recorro todos los posibles modelos a probar en iteracion j del outter loop\n",
        "          grid_search = GridSearchCV(Pipeline(steps[r],memory=pathpre), parameters[r],cv=5,verbose=5,scoring='neg_mean_absolute_error',n_jobs=-1) #cv inner loop\n",
        "          #xtrain gridsearchcv xtrain split en 12 / cv, 60/5 = 12, xtrain 48 datos validar 12\n",
        "          # cv = N -> leave one out N <30\n",
        "          #generar mejor modelo\n",
        "          grid_search.fit(X_trainj,y_trainj)\n",
        "          #estimar salida conjunto de test\n",
        "          y_pred = grid_search.best_estimator_.predict(X_testj)\n",
        "          #guardar mejor modelo\n",
        "          list_est.append(grid_search.best_estimator_)\n",
        "          list_hyper.append(grid_search.best_params_)\n",
        "          #guardar acierto\n",
        "          msev[j,r] = msa(y_testj,y_pred)\n",
        "          print('it %d/%d-Modelo %d/%d'%(j+1,Niter,r+1,len(steps)))\n",
        "          print('best hyper', grid_search.best_params_)\n",
        "          print('msa:',msev[j,r])\n",
        "          \n",
        "      best_estimators[j] = list_est #guardar mejores modelos \n",
        "      best_hyperpar[j] = list_hyper #mejores hyperparametros\n",
        "          \n",
        "          \n",
        "      savedata = {\n",
        "          'acc':msev,\n",
        "          'best_models':best_estimators,\n",
        "          'best_parameters':best_hyperpar,\n",
        "            } \n",
        "      dump(savedata,'videogames.joblib')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 5 folds for each of 2 candidates, totalling 10 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    1.5s remaining:    0.0s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    1.5s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 1/10-Modelo 1/2\n",
            "best hyper {'reg__fit_intercept': True}\n",
            "msa: 4.011455554847524\n",
            "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  24 tasks      | elapsed:    2.7s\n",
            "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:    7.8s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 1/10-Modelo 2/2\n",
            "best hyper {'reg__alpha': 0.01, 'reg__l1_ratio': 0.5}\n",
            "msa: 4.012725981438147\n",
            "Fitting 5 folds for each of 2 candidates, totalling 10 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s remaining:    0.0s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 2/10-Modelo 1/2\n",
            "best hyper {'reg__fit_intercept': True}\n",
            "msa: 4.120048046121391\n",
            "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  14 tasks      | elapsed:    1.8s\n",
            "[Parallel(n_jobs=-1)]: Done  70 tasks      | elapsed:    6.2s\n",
            "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:    8.6s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 2/10-Modelo 2/2\n",
            "best hyper {'reg__alpha': 0, 'reg__l1_ratio': 0}\n",
            "msa: 4.119880181108407\n",
            "Fitting 5 folds for each of 2 candidates, totalling 10 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done   7 out of  10 | elapsed:    0.3s remaining:    0.1s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s remaining:    0.0s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 3/10-Modelo 1/2\n",
            "best hyper {'reg__fit_intercept': True}\n",
            "msa: 4.0580593613251255\n",
            "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  24 tasks      | elapsed:    2.6s\n",
            "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:    7.8s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 3/10-Modelo 2/2\n",
            "best hyper {'reg__alpha': 0, 'reg__l1_ratio': 0}\n",
            "msa: 4.057906966810335\n",
            "Fitting 5 folds for each of 2 candidates, totalling 10 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s remaining:    0.0s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 4/10-Modelo 1/2\n",
            "best hyper {'reg__fit_intercept': True}\n",
            "msa: 4.039188287549976\n",
            "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  14 tasks      | elapsed:    1.4s\n",
            "[Parallel(n_jobs=-1)]: Done  70 tasks      | elapsed:    5.4s\n",
            "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:    7.8s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 4/10-Modelo 2/2\n",
            "best hyper {'reg__alpha': 0, 'reg__l1_ratio': 0}\n",
            "msa: 4.03854936427749\n",
            "Fitting 5 folds for each of 2 candidates, totalling 10 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done   7 out of  10 | elapsed:    0.3s remaining:    0.1s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s remaining:    0.0s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 5/10-Modelo 1/2\n",
            "best hyper {'reg__fit_intercept': True}\n",
            "msa: 4.099079033387267\n",
            "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  24 tasks      | elapsed:    2.7s\n",
            "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:    7.8s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 5/10-Modelo 2/2\n",
            "best hyper {'reg__alpha': 0, 'reg__l1_ratio': 0}\n",
            "msa: 4.099052681517665\n",
            "Fitting 5 folds for each of 2 candidates, totalling 10 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done   7 out of  10 | elapsed:    0.3s remaining:    0.1s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s remaining:    0.0s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 6/10-Modelo 1/2\n",
            "best hyper {'reg__fit_intercept': True}\n",
            "msa: 4.069386180102924\n",
            "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  24 tasks      | elapsed:    2.3s\n",
            "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:    7.6s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 6/10-Modelo 2/2\n",
            "best hyper {'reg__alpha': 0, 'reg__l1_ratio': 0}\n",
            "msa: 4.068440434125986\n",
            "Fitting 5 folds for each of 2 candidates, totalling 10 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done   7 out of  10 | elapsed:    0.3s remaining:    0.1s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s remaining:    0.0s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 7/10-Modelo 1/2\n",
            "best hyper {'reg__fit_intercept': True}\n",
            "msa: 4.048851591100405\n",
            "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  14 tasks      | elapsed:    1.6s\n",
            "[Parallel(n_jobs=-1)]: Done  70 tasks      | elapsed:    6.0s\n",
            "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:    8.4s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 7/10-Modelo 2/2\n",
            "best hyper {'reg__alpha': 0.01, 'reg__l1_ratio': 1}\n",
            "msa: 4.048270903172693\n",
            "Fitting 5 folds for each of 2 candidates, totalling 10 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done   7 out of  10 | elapsed:    0.3s remaining:    0.1s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.4s remaining:    0.0s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.4s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 8/10-Modelo 1/2\n",
            "best hyper {'reg__fit_intercept': True}\n",
            "msa: 4.080891929371928\n",
            "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  24 tasks      | elapsed:    2.5s\n",
            "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:    7.3s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 8/10-Modelo 2/2\n",
            "best hyper {'reg__alpha': 0, 'reg__l1_ratio': 0}\n",
            "msa: 4.080204228900644\n",
            "Fitting 5 folds for each of 2 candidates, totalling 10 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done   7 out of  10 | elapsed:    0.3s remaining:    0.1s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s remaining:    0.0s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 9/10-Modelo 1/2\n",
            "best hyper {'reg__fit_intercept': True}\n",
            "msa: 4.095596525389448\n",
            "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  24 tasks      | elapsed:    2.4s\n",
            "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:    7.4s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 9/10-Modelo 2/2\n",
            "best hyper {'reg__alpha': 0, 'reg__l1_ratio': 0}\n",
            "msa: 4.0947884060001725\n",
            "Fitting 5 folds for each of 2 candidates, totalling 10 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done   7 out of  10 | elapsed:    0.3s remaining:    0.1s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s remaining:    0.0s\n",
            "[Parallel(n_jobs=-1)]: Done  10 out of  10 | elapsed:    0.3s finished\n",
            "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 2 concurrent workers.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 10/10-Modelo 1/2\n",
            "best hyper {'reg__fit_intercept': True}\n",
            "msa: 4.078275872400066\n",
            "Fitting 5 folds for each of 30 candidates, totalling 150 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done  24 tasks      | elapsed:    2.8s\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "it 10/10-Modelo 2/2\n",
            "best hyper {'reg__alpha': 0.01, 'reg__l1_ratio': 1}\n",
            "msa: 4.0787144163712\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=-1)]: Done 150 out of 150 | elapsed:    7.9s finished\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1ppoKkgal9Bt",
        "outputId": "77ad4c06-43b5-4f3e-e253-4e2c6a0c66b3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        }
      },
      "source": [
        "from scipy.stats import mode\n",
        "from datetime import date\n",
        "\n",
        "r = 1 #hyperparameters model 1\n",
        "nh = len(best_hyperpar[0][r])\n",
        "hyperpar_r = np.zeros((Niter,nh)) \n",
        "for i in range(Niter):\n",
        "  for j in range(nh):\n",
        "    hyperpar_r[i,j] = best_hyperpar[i][r].get(list(best_hyperpar[i][r].keys())[j])\n",
        "    \n",
        "    \n",
        "#revisar numero entero para realizar casting\n",
        "aa = list(best_hyperpar[0][r].keys())\n",
        "c = -1\n",
        "for i in range(len(aa)):\n",
        "    if aa[i].find('n_neighbors') > -1:\n",
        "      c = i\n",
        "\n",
        "plt.boxplot(hyperpar_r)\n",
        "plt.xticks(ticks=np.arange(nh)+1,labels=list(best_hyperpar[0][r].keys()))\n",
        "plt.title('Best_hyperparameters '+label_model[r])\n",
        "plt.grid()\n",
        "plt.show()\n",
        "\n",
        "mode_hyper = mode(hyperpar_r,axis=0)[0][0]\n",
        "print(\"Modes= \", mode_hyper)"
      ],
      "execution_count": null,
      "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"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Modes=  [0. 0.]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "toLYISEVmTKp"
      },
      "source": [
        "# Modelo definitivo\n",
        "\n",
        "Con los mejores hiperparámetros entreno sobre Xtrain completo el modelo final"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_NoVVCBTmPje",
        "outputId": "cd1a83e0-5143-4e4f-b569-e2178e757f83",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 195
        }
      },
      "source": [
        "r = 1 #camino elastic net\n",
        "steps_final = [('scaler',StandardScaler()),\n",
        "               ('reg',ElasticNet(alpha=0.0,l1_ratio=0))]  \n",
        "         \n",
        "modelo_final = Pipeline(steps_final)\n",
        "modelo_final.fit(dftrain_pre,ytrain)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Pipeline(memory=None,\n",
              "         steps=[('scaler',\n",
              "                 StandardScaler(copy=True, with_mean=True, with_std=True)),\n",
              "                ('reg',\n",
              "                 ElasticNet(alpha=0.0, copy_X=True, fit_intercept=True,\n",
              "                            l1_ratio=0, max_iter=1000, normalize=False,\n",
              "                            positive=False, precompute=False, random_state=None,\n",
              "                            selection='cyclic', tol=0.0001,\n",
              "                            warm_start=False))],\n",
              "         verbose=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "O4M6kYhRm28a"
      },
      "source": [
        "# Caso real evaluación datos nuevos"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UKWUcLFIm3-7"
      },
      "source": [
        "ytest_e = modelo_final.predict(dummy.transform(df_test)) #simular casos nuevos"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ymoZ1E15m9Hl",
        "outputId": "b372204e-b9ea-4da0-975a-8b1e08125085",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "print('MAE_test=', msa(ytest,ytest_e))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "MAE_test= 3.9567832730288717\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7r2zZSIPm9hr",
        "outputId": "269822ea-f903-4369-e042-76d11f0c09fd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "pathpre = 'resultados'\n",
        "\n",
        "try:\n",
        "  os.mkdir(pathpre)\n",
        "except:\n",
        "  print(\"Carpeta results ya existe\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Carpeta results ya existe\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UQTAZOSym_OJ",
        "outputId": "254a3a78-ffb9-4b43-9e98-7a9e28baf00f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 17
        }
      },
      "source": [
        "import shutil\n",
        "from joblib import dump, load\n",
        "from datetime import date, datetime\n",
        "from google.colab import files\n",
        "\n",
        "modelo_final = {'modelo':modelo_final,\n",
        "          'pasos':label_model[1],\n",
        "          'mae_test':msa(ytest,ytest_e),\n",
        "            } \n",
        "\n",
        "dump(modelo_final,'resultados/modelo_final.joblib')\n",
        "namefile = str(datetime.now().strftime(\"%Y_%m_%d_%H_%M_%d\"))+'__modelo'\n",
        "shutil.make_archive(namefile, 'zip', 'resultados')\n",
        "files.download(namefile+'.zip')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_b7f2dd1d-35f6-4676-96df-6febedadb661\", \"2020_10_21_00_06_21__modelo.zip\", 1502)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AaAwpk00nBC_"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}