E08_SentimentPrediction.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "E08-SentimentPrediction.ipynb",
      "version": "0.3.2",
      "views": {},
      "default_view": {},
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    }
  },
  "cells": [
    {
      "metadata": {
        "id": "aog2MshfrKKq",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "\n",
        "# Deep Learning Course\n",
        "\n",
        " <img src=\"https://d2h0cx97tjks2p.cloudfront.net/blogs/wp-content/uploads/Natural-Language-Processing.jpg\" >\n",
        "\n",
        "\n",
        "# Deep Learning Excercise 08\n",
        "## Analyze how travelers expressed their feelings on Twitter\n",
        "\n",
        "\n",
        "Developed by [Sergio Castelblanco](http://www.linkedin.com/in/sergio-castelblanco/) - [Jhon Gonzalez](http://linkedin.com/in/jhon-franklin-gonzalez-gamboa-b6123230/)\n",
        "\n",
        "Andes University\n",
        "\n",
        "version 2.0, July 18 2018"
      ]
    },
    {
      "metadata": {
        "id": "LqX4zdJ4rKKr",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "A sentiment analysis job about the problems of each major U.S. airline. \n",
        "Twitter data was scraped from February of 2015 and contributors were \n",
        "asked to first classify positive, negative, and neutral tweets, followed\n",
        "by categorizing negative reasons (such as \"late flight\" or \"rude service\")."
      ]
    },
    {
      "metadata": {
        "id": "CXugBIpUrKKr",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Our target is to classify a feeling tweet message feelings into three possible categories:\n",
        "\n",
        "* Positive\n",
        "* Neural\n",
        "* Negative\n",
        "\n",
        "We will divide the dataset into train and test and we will recognize the testing tweet feelings\n"
      ]
    },
    {
      "metadata": {
        "id": "EyNHGVLNrKKs",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "90910724-2fb3-41a5-ec87-554ab87cfe00"
      },
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "\n",
        "%matplotlib inline\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# read the data and set the datetime as the index\n",
        "import zipfile\n",
        "with zipfile.ZipFile('../datasets/Tweets.zip', 'r') as z:\n",
        "    f = z.open('Tweets.csv')\n",
        "    tweets = pd.read_csv(f, index_col=0)\n",
        "\n",
        "tweets.head()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style>\n",
              "    .dataframe thead tr:only-child th {\n",
              "        text-align: right;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: left;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>airline_sentiment</th>\n",
              "      <th>airline_sentiment_confidence</th>\n",
              "      <th>negativereason</th>\n",
              "      <th>negativereason_confidence</th>\n",
              "      <th>airline</th>\n",
              "      <th>airline_sentiment_gold</th>\n",
              "      <th>name</th>\n",
              "      <th>negativereason_gold</th>\n",
              "      <th>retweet_count</th>\n",
              "      <th>text</th>\n",
              "      <th>tweet_coord</th>\n",
              "      <th>tweet_created</th>\n",
              "      <th>tweet_location</th>\n",
              "      <th>user_timezone</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tweet_id</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>570306133677760513</th>\n",
              "      <td>neutral</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Virgin America</td>\n",
              "      <td>NaN</td>\n",
              "      <td>cairdin</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "      <td>@VirginAmerica What @dhepburn said.</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2015-02-24 11:35:52 -0800</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Eastern Time (US &amp; Canada)</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570301130888122368</th>\n",
              "      <td>positive</td>\n",
              "      <td>0.3486</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>Virgin America</td>\n",
              "      <td>NaN</td>\n",
              "      <td>jnardino</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2015-02-24 11:15:59 -0800</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Pacific Time (US &amp; Canada)</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570301083672813571</th>\n",
              "      <td>neutral</td>\n",
              "      <td>0.6837</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Virgin America</td>\n",
              "      <td>NaN</td>\n",
              "      <td>yvonnalynn</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2015-02-24 11:15:48 -0800</td>\n",
              "      <td>Lets Play</td>\n",
              "      <td>Central Time (US &amp; Canada)</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570301031407624196</th>\n",
              "      <td>negative</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>Bad Flight</td>\n",
              "      <td>0.7033</td>\n",
              "      <td>Virgin America</td>\n",
              "      <td>NaN</td>\n",
              "      <td>jnardino</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2015-02-24 11:15:36 -0800</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Pacific Time (US &amp; Canada)</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570300817074462722</th>\n",
              "      <td>negative</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>Can't Tell</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>Virgin America</td>\n",
              "      <td>NaN</td>\n",
              "      <td>jnardino</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2015-02-24 11:14:45 -0800</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Pacific Time (US &amp; Canada)</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                   airline_sentiment  airline_sentiment_confidence  \\\n",
              "tweet_id                                                             \n",
              "570306133677760513           neutral                        1.0000   \n",
              "570301130888122368          positive                        0.3486   \n",
              "570301083672813571           neutral                        0.6837   \n",
              "570301031407624196          negative                        1.0000   \n",
              "570300817074462722          negative                        1.0000   \n",
              "\n",
              "                   negativereason  negativereason_confidence         airline  \\\n",
              "tweet_id                                                                       \n",
              "570306133677760513            NaN                        NaN  Virgin America   \n",
              "570301130888122368            NaN                     0.0000  Virgin America   \n",
              "570301083672813571            NaN                        NaN  Virgin America   \n",
              "570301031407624196     Bad Flight                     0.7033  Virgin America   \n",
              "570300817074462722     Can't Tell                     1.0000  Virgin America   \n",
              "\n",
              "                   airline_sentiment_gold        name negativereason_gold  \\\n",
              "tweet_id                                                                    \n",
              "570306133677760513                    NaN     cairdin                 NaN   \n",
              "570301130888122368                    NaN    jnardino                 NaN   \n",
              "570301083672813571                    NaN  yvonnalynn                 NaN   \n",
              "570301031407624196                    NaN    jnardino                 NaN   \n",
              "570300817074462722                    NaN    jnardino                 NaN   \n",
              "\n",
              "                    retweet_count  \\\n",
              "tweet_id                            \n",
              "570306133677760513              0   \n",
              "570301130888122368              0   \n",
              "570301083672813571              0   \n",
              "570301031407624196              0   \n",
              "570300817074462722              0   \n",
              "\n",
              "                                                                 text  \\\n",
              "tweet_id                                                                \n",
              "570306133677760513                @VirginAmerica What @dhepburn said.   \n",
              "570301130888122368  @VirginAmerica plus you've added commercials t...   \n",
              "570301083672813571  @VirginAmerica I didn't today... Must mean I n...   \n",
              "570301031407624196  @VirginAmerica it's really aggressive to blast...   \n",
              "570300817074462722  @VirginAmerica and it's a really big bad thing...   \n",
              "\n",
              "                   tweet_coord              tweet_created tweet_location  \\\n",
              "tweet_id                                                                   \n",
              "570306133677760513         NaN  2015-02-24 11:35:52 -0800            NaN   \n",
              "570301130888122368         NaN  2015-02-24 11:15:59 -0800            NaN   \n",
              "570301083672813571         NaN  2015-02-24 11:15:48 -0800      Lets Play   \n",
              "570301031407624196         NaN  2015-02-24 11:15:36 -0800            NaN   \n",
              "570300817074462722         NaN  2015-02-24 11:14:45 -0800            NaN   \n",
              "\n",
              "                                 user_timezone  \n",
              "tweet_id                                        \n",
              "570306133677760513  Eastern Time (US & Canada)  \n",
              "570301130888122368  Pacific Time (US & Canada)  \n",
              "570301083672813571  Central Time (US & Canada)  \n",
              "570301031407624196  Pacific Time (US & Canada)  \n",
              "570300817074462722  Pacific Time (US & Canada)  "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    },
    {
      "metadata": {
        "id": "IUxremfBrKKx",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "11b0c938-6537-4c34-cbc0-375245a36d63"
      },
      "cell_type": "code",
      "source": [
        "tweets.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(14640, 14)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "metadata": {
        "id": "0e2UUW2KrKK0",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Proportion of tweets with each sentiment"
      ]
    },
    {
      "metadata": {
        "id": "gr2H2taQrKK0",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "cd638f28-f385-4824-d460-92953fb224d0"
      },
      "cell_type": "code",
      "source": [
        "tweets['airline_sentiment'].value_counts()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "negative    9178\n",
              "neutral     3099\n",
              "positive    2363\n",
              "Name: airline_sentiment, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "metadata": {
        "id": "N5-2mUxwrKK2",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Proportion of tweets per airline\n"
      ]
    },
    {
      "metadata": {
        "id": "wF7xgO72rKK3",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "5d676dab-9eb4-41f1-b7af-5842f41ea11f"
      },
      "cell_type": "code",
      "source": [
        "tweets['airline'].value_counts()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "United            3822\n",
              "US Airways        2913\n",
              "American          2759\n",
              "Southwest         2420\n",
              "Delta             2222\n",
              "Virgin America     504\n",
              "Name: airline, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "metadata": {
        "id": "hA5d_lEPrKK6",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "b34b60b6-f27f-43f5-f3ec-8a3ffc846da0"
      },
      "cell_type": "code",
      "source": [
        "tweets[['airline_sentiment', 'text']]"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style>\n",
              "    .dataframe thead tr:only-child th {\n",
              "        text-align: right;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: left;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>airline_sentiment</th>\n",
              "      <th>text</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tweet_id</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>570306133677760513</th>\n",
              "      <td>neutral</td>\n",
              "      <td>@VirginAmerica What @dhepburn said.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570301130888122368</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica plus you've added commercials t...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570301083672813571</th>\n",
              "      <td>neutral</td>\n",
              "      <td>@VirginAmerica I didn't today... Must mean I n...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570301031407624196</th>\n",
              "      <td>negative</td>\n",
              "      <td>@VirginAmerica it's really aggressive to blast...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570300817074462722</th>\n",
              "      <td>negative</td>\n",
              "      <td>@VirginAmerica and it's a really big bad thing...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570300767074181121</th>\n",
              "      <td>negative</td>\n",
              "      <td>@VirginAmerica seriously would pay $30 a fligh...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570300616901320704</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica yes, nearly every time I fly VX...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570300248553349120</th>\n",
              "      <td>neutral</td>\n",
              "      <td>@VirginAmerica Really missed a prime opportuni...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570299953286942721</th>\n",
              "      <td>positive</td>\n",
              "      <td>@virginamerica Well, I didn't…but NOW I DO! :-D</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570295459631263746</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica it was amazing, and arrived an ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570294189143031808</th>\n",
              "      <td>neutral</td>\n",
              "      <td>@VirginAmerica did you know that suicide is th...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570289724453216256</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica I &amp;lt;3 pretty graphics. so muc...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570289584061480960</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica This is such a great deal! Alre...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570287408438120448</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica @virginmedia I'm flying your #f...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570285904809598977</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica Thanks!</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570282469121007616</th>\n",
              "      <td>negative</td>\n",
              "      <td>@VirginAmerica SFO-PDX schedule is still MIA.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570277724385734656</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica So excited for my first cross c...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570276917301137409</th>\n",
              "      <td>negative</td>\n",
              "      <td>@VirginAmerica  I flew from NYC to SFO last we...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570270684619923457</th>\n",
              "      <td>positive</td>\n",
              "      <td>I ❤️ flying @VirginAmerica. ☺️👍</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570267956648792064</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica you know what would be amazingl...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570265883513384960</th>\n",
              "      <td>negative</td>\n",
              "      <td>@VirginAmerica why are your first fares in May...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570264145116819457</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica I love this graphic. http://t.c...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570259420287868928</th>\n",
              "      <td>positive</td>\n",
              "      <td>@VirginAmerica I love the hipster innovation. ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570258822297579520</th>\n",
              "      <td>neutral</td>\n",
              "      <td>@VirginAmerica will you be making BOS&amp;gt;LAS n...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570256553502068736</th>\n",
              "      <td>negative</td>\n",
              "      <td>@VirginAmerica you guys messed up my seating.....</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570249102404923392</th>\n",
              "      <td>negative</td>\n",
              "      <td>@VirginAmerica status match program.  I applie...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570239632807370753</th>\n",
              "      <td>negative</td>\n",
              "      <td>@VirginAmerica What happened 2 ur vegan food o...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570217831557677057</th>\n",
              "      <td>neutral</td>\n",
              "      <td>@VirginAmerica do you miss me? Don't worry we'...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570207886493782019</th>\n",
              "      <td>negative</td>\n",
              "      <td>@VirginAmerica amazing to me that we can't get...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>570124596180955136</th>\n",
              "      <td>neutral</td>\n",
              "      <td>@VirginAmerica LAX to EWR - Middle seat on a r...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569591765793165312</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir I understand the weather issue bu...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569591730506371072</th>\n",
              "      <td>neutral</td>\n",
              "      <td>@AmericanAir guarantee no retribution? If so, ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569591700416393216</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir a friend is having flight Cancell...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569591653121597440</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir I used the \"call back\" feature wi...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569591540944756737</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir I need to be at work tomorrow at ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569591533617307648</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir  ugh Dump us in dfw w/no luggage ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569591393540288512</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir Cancelled Flights my flight, does...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569591285150908416</th>\n",
              "      <td>positive</td>\n",
              "      <td>@AmericanAir DMing you now! Big thanks.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569591136534319105</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir 3078 is overweight so you pull 2 ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569590988395708416</th>\n",
              "      <td>positive</td>\n",
              "      <td>@AmericanAir I love your company and your staf...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569590965880532993</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir I wait 2+ hrs for CS to call me b...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569590892085915649</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir I've been on hold for 55 mins abo...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569590191758962688</th>\n",
              "      <td>negative</td>\n",
              "      <td>I just need a place to sleep when I land witho...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569590013278756865</th>\n",
              "      <td>positive</td>\n",
              "      <td>@AmericanAir Love the new planes for the JFK-L...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569589959088173056</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir Call me Chairman, or call me Emer...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569589643487928321</th>\n",
              "      <td>positive</td>\n",
              "      <td>@AmericanAir Flight 236 was great. Fantastic c...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569589460226183168</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir Flight 953 NYC-Buenos Aires has b...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569588816438169600</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir Flight Cancelled Flightled, can't...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569588651925098496</th>\n",
              "      <td>positive</td>\n",
              "      <td>Thank you. “@AmericanAir: @jlhalldc Customer R...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569588591602458624</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir How do I change my flight if the ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569588473050611712</th>\n",
              "      <td>positive</td>\n",
              "      <td>@AmericanAir Thanks! He is.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569588464896876545</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir thx for nothing on getting us out...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569587813856841728</th>\n",
              "      <td>neutral</td>\n",
              "      <td>“@AmericanAir: @TilleyMonsta George, that does...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569587705937600512</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir my flight was Cancelled Flightled...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569587691626622976</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir right on cue with the delays👌</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569587686496825344</th>\n",
              "      <td>positive</td>\n",
              "      <td>@AmericanAir thank you we got on a different f...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569587371693355008</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir leaving over 20 minutes Late Flig...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569587242672398336</th>\n",
              "      <td>neutral</td>\n",
              "      <td>@AmericanAir Please bring American Airlines to...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569587188687634433</th>\n",
              "      <td>negative</td>\n",
              "      <td>@AmericanAir you have my money, you change my ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>569587140490866689</th>\n",
              "      <td>neutral</td>\n",
              "      <td>@AmericanAir we have 8 ppl so we need 2 know h...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>14640 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                   airline_sentiment  \\\n",
              "tweet_id                               \n",
              "570306133677760513           neutral   \n",
              "570301130888122368          positive   \n",
              "570301083672813571           neutral   \n",
              "570301031407624196          negative   \n",
              "570300817074462722          negative   \n",
              "570300767074181121          negative   \n",
              "570300616901320704          positive   \n",
              "570300248553349120           neutral   \n",
              "570299953286942721          positive   \n",
              "570295459631263746          positive   \n",
              "570294189143031808           neutral   \n",
              "570289724453216256          positive   \n",
              "570289584061480960          positive   \n",
              "570287408438120448          positive   \n",
              "570285904809598977          positive   \n",
              "570282469121007616          negative   \n",
              "570277724385734656          positive   \n",
              "570276917301137409          negative   \n",
              "570270684619923457          positive   \n",
              "570267956648792064          positive   \n",
              "570265883513384960          negative   \n",
              "570264145116819457          positive   \n",
              "570259420287868928          positive   \n",
              "570258822297579520           neutral   \n",
              "570256553502068736          negative   \n",
              "570249102404923392          negative   \n",
              "570239632807370753          negative   \n",
              "570217831557677057           neutral   \n",
              "570207886493782019          negative   \n",
              "570124596180955136           neutral   \n",
              "...                              ...   \n",
              "569591765793165312          negative   \n",
              "569591730506371072           neutral   \n",
              "569591700416393216          negative   \n",
              "569591653121597440          negative   \n",
              "569591540944756737          negative   \n",
              "569591533617307648          negative   \n",
              "569591393540288512          negative   \n",
              "569591285150908416          positive   \n",
              "569591136534319105          negative   \n",
              "569590988395708416          positive   \n",
              "569590965880532993          negative   \n",
              "569590892085915649          negative   \n",
              "569590191758962688          negative   \n",
              "569590013278756865          positive   \n",
              "569589959088173056          negative   \n",
              "569589643487928321          positive   \n",
              "569589460226183168          negative   \n",
              "569588816438169600          negative   \n",
              "569588651925098496          positive   \n",
              "569588591602458624          negative   \n",
              "569588473050611712          positive   \n",
              "569588464896876545          negative   \n",
              "569587813856841728           neutral   \n",
              "569587705937600512          negative   \n",
              "569587691626622976          negative   \n",
              "569587686496825344          positive   \n",
              "569587371693355008          negative   \n",
              "569587242672398336           neutral   \n",
              "569587188687634433          negative   \n",
              "569587140490866689           neutral   \n",
              "\n",
              "                                                                 text  \n",
              "tweet_id                                                               \n",
              "570306133677760513                @VirginAmerica What @dhepburn said.  \n",
              "570301130888122368  @VirginAmerica plus you've added commercials t...  \n",
              "570301083672813571  @VirginAmerica I didn't today... Must mean I n...  \n",
              "570301031407624196  @VirginAmerica it's really aggressive to blast...  \n",
              "570300817074462722  @VirginAmerica and it's a really big bad thing...  \n",
              "570300767074181121  @VirginAmerica seriously would pay $30 a fligh...  \n",
              "570300616901320704  @VirginAmerica yes, nearly every time I fly VX...  \n",
              "570300248553349120  @VirginAmerica Really missed a prime opportuni...  \n",
              "570299953286942721    @virginamerica Well, I didn't…but NOW I DO! :-D  \n",
              "570295459631263746  @VirginAmerica it was amazing, and arrived an ...  \n",
              "570294189143031808  @VirginAmerica did you know that suicide is th...  \n",
              "570289724453216256  @VirginAmerica I &lt;3 pretty graphics. so muc...  \n",
              "570289584061480960  @VirginAmerica This is such a great deal! Alre...  \n",
              "570287408438120448  @VirginAmerica @virginmedia I'm flying your #f...  \n",
              "570285904809598977                             @VirginAmerica Thanks!  \n",
              "570282469121007616      @VirginAmerica SFO-PDX schedule is still MIA.  \n",
              "570277724385734656  @VirginAmerica So excited for my first cross c...  \n",
              "570276917301137409  @VirginAmerica  I flew from NYC to SFO last we...  \n",
              "570270684619923457                    I ❤️ flying @VirginAmerica. ☺️👍  \n",
              "570267956648792064  @VirginAmerica you know what would be amazingl...  \n",
              "570265883513384960  @VirginAmerica why are your first fares in May...  \n",
              "570264145116819457  @VirginAmerica I love this graphic. http://t.c...  \n",
              "570259420287868928  @VirginAmerica I love the hipster innovation. ...  \n",
              "570258822297579520  @VirginAmerica will you be making BOS&gt;LAS n...  \n",
              "570256553502068736  @VirginAmerica you guys messed up my seating.....  \n",
              "570249102404923392  @VirginAmerica status match program.  I applie...  \n",
              "570239632807370753  @VirginAmerica What happened 2 ur vegan food o...  \n",
              "570217831557677057  @VirginAmerica do you miss me? Don't worry we'...  \n",
              "570207886493782019  @VirginAmerica amazing to me that we can't get...  \n",
              "570124596180955136  @VirginAmerica LAX to EWR - Middle seat on a r...  \n",
              "...                                                               ...  \n",
              "569591765793165312  @AmericanAir I understand the weather issue bu...  \n",
              "569591730506371072  @AmericanAir guarantee no retribution? If so, ...  \n",
              "569591700416393216  @AmericanAir a friend is having flight Cancell...  \n",
              "569591653121597440  @AmericanAir I used the \"call back\" feature wi...  \n",
              "569591540944756737  @AmericanAir I need to be at work tomorrow at ...  \n",
              "569591533617307648  @AmericanAir  ugh Dump us in dfw w/no luggage ...  \n",
              "569591393540288512  @AmericanAir Cancelled Flights my flight, does...  \n",
              "569591285150908416            @AmericanAir DMing you now! Big thanks.  \n",
              "569591136534319105  @AmericanAir 3078 is overweight so you pull 2 ...  \n",
              "569590988395708416  @AmericanAir I love your company and your staf...  \n",
              "569590965880532993  @AmericanAir I wait 2+ hrs for CS to call me b...  \n",
              "569590892085915649  @AmericanAir I've been on hold for 55 mins abo...  \n",
              "569590191758962688  I just need a place to sleep when I land witho...  \n",
              "569590013278756865  @AmericanAir Love the new planes for the JFK-L...  \n",
              "569589959088173056  @AmericanAir Call me Chairman, or call me Emer...  \n",
              "569589643487928321  @AmericanAir Flight 236 was great. Fantastic c...  \n",
              "569589460226183168  @AmericanAir Flight 953 NYC-Buenos Aires has b...  \n",
              "569588816438169600  @AmericanAir Flight Cancelled Flightled, can't...  \n",
              "569588651925098496  Thank you. “@AmericanAir: @jlhalldc Customer R...  \n",
              "569588591602458624  @AmericanAir How do I change my flight if the ...  \n",
              "569588473050611712                        @AmericanAir Thanks! He is.  \n",
              "569588464896876545  @AmericanAir thx for nothing on getting us out...  \n",
              "569587813856841728  “@AmericanAir: @TilleyMonsta George, that does...  \n",
              "569587705937600512  @AmericanAir my flight was Cancelled Flightled...  \n",
              "569587691626622976         @AmericanAir right on cue with the delays👌  \n",
              "569587686496825344  @AmericanAir thank you we got on a different f...  \n",
              "569587371693355008  @AmericanAir leaving over 20 minutes Late Flig...  \n",
              "569587242672398336  @AmericanAir Please bring American Airlines to...  \n",
              "569587188687634433  @AmericanAir you have my money, you change my ...  \n",
              "569587140490866689  @AmericanAir we have 8 ppl so we need 2 know h...  \n",
              "\n",
              "[14640 rows x 2 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "metadata": {
        "id": "UPmvIO_KrKK9",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "cdc42563-85a3-45bc-8aae-d32bdd129163"
      },
      "cell_type": "code",
      "source": [
        "pd.Series(tweets[\"airline\"]).value_counts().plot(kind = \"bar\",figsize=(8,6),rot = 0)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x117193be0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfAAAAFpCAYAAABjxXptAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n9P8ByH7UwFdC2IIAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x117193f28>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "Go0CCBTxrKLA",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "eead8f60-e888-4759-e318-99ffa2342317"
      },
      "cell_type": "code",
      "source": [
        "pd.crosstab(index = tweets[\"airline\"],columns = tweets[\"airline_sentiment\"]).plot(kind='bar',figsize=(10, 6),alpha=0.5,rot=0,stacked=True,title=\"Sentiment by airline\")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x115f9abe0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmAAAAGDCAYAAACMU6xhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nklSx/w8tJgxPCiM0LAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x1125df358>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "Wx4XhLSYrKLF",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Exercise 8.1 \n",
        "\n",
        "Predict the sentiment using CountVectorizer, stopwords, n_grams, stemmer, TfidfVectorizer\n",
        "\n",
        "use Random Forest classifier"
      ]
    },
    {
      "metadata": {
        "id": "m7-317TUrKLG",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "# ya está partido en X y Y, crear un Random Forest para predecir el sentimiento de los Tweets\n",
        "# count vectorizer, hacer el flujo como el notebook 19."
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "RoQtpxyhrKLI",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split, cross_val_score\n",
        "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from nltk.stem.snowball import SnowballStemmer\n",
        "from nltk.stem import WordNetLemmatizer"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "_9C8JBKdrKLK",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "X = tweets['text']\n",
        "y = tweets['airline_sentiment'].map({'negative':-1,'neutral':0,'positive':1})"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Yu0sOw0krKLO",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "79a8736f-fa0c-45b2-d2a4-8c04327ce173"
      },
      "cell_type": "code",
      "source": [
        "y.describe()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "count    14640.000000\n",
              "mean        -0.465505\n",
              "std          0.756084\n",
              "min         -1.000000\n",
              "25%         -1.000000\n",
              "50%         -1.000000\n",
              "75%          0.000000\n",
              "max          1.000000\n",
              "Name: airline_sentiment, dtype: float64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "metadata": {
        "id": "6pwVh3AUrKLR",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "03fbb20c-9e03-4811-86e8-db26a2882784"
      },
      "cell_type": "code",
      "source": [
        "from keras.utils import np_utils\n",
        "y_train= np_utils.to_categorical(y, 3)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "metadata": {
        "id": "IIF80idQrKLT",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "We will split our data into train a test"
      ]
    },
    {
      "metadata": {
        "id": "FEkP-sItrKLU",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y_train, test_size=0.33, random_state=42)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "oJC4h7iPrKLX",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "885ac3b8-d389-4d15-e5f9-08b6c871929a"
      },
      "cell_type": "code",
      "source": [
        "vect = CountVectorizer()\n",
        "X_dtm = vect.fit_transform(X_train)\n",
        "X_dtm.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(9808, 12056)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "metadata": {
        "id": "e1PDcr9irKLa",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "96560a04-d970-4945-f549-a413404c0a99"
      },
      "cell_type": "code",
      "source": [
        "print(vect.get_feature_names()[-150:-100])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['xc6jq70r7b', 'xcvqxykg49', 'xd', 'xdjzkc34gb', 'xedeckgmw5', 'xf9mpnaaje', 'xgfs6tjtmo', 'xgpaayfdwt', 'xhsuupheze', 'xijifymvqp', 'xm', 'xmrvr4lgeg', 'xmz3tf9ix8', 'xna', 'xox', 'xoxo', 'xpm98igqjn', 'xrdtov7nl8', 'xtra', 'xx', 'xx2m2jxqep', 'xx5qscjll1', 'xxppzo88j1', 'xxsaombdgs', 'xxy2d2imnp', 'xzbajmiekx', 'xzmscw', 'y3sahvx3zk', 'y5qe9hcqzt', 'ya', 'yaffasolin', 'yall', 'yards', 'yasss', 'yasssss', 'yay', 'yayayay', 'ybmbgs0dhn', 'yc7v2s0iod', 'yco9dikpt9', 'yday', 'yea', 'yeah', 'year', 'yearly', 'years', 'yeehaw', 'yeg', 'yell', 'yelled']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "r7FZP1ZlrKLc",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "a51deb6a-6eed-4742-be90-fa3bf2e6fc92"
      },
      "cell_type": "code",
      "source": [
        "vect = CountVectorizer(ngram_range=(1, 2))\n",
        "X_dtm = vect.fit_transform(X_train)\n",
        "X_dtm.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(9808, 87608)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "metadata": {
        "id": "g0faqthXrKLf",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "27180833-28f3-4fe4-d97b-69cf439e22c2"
      },
      "cell_type": "code",
      "source": [
        "X_dtm.todense()[0].argmax()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "6489"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "metadata": {
        "id": "fDOqro0jrKLh",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "The most common word is am on the training data set"
      ]
    },
    {
      "metadata": {
        "id": "R0qU3s0grKLi",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "06985fa9-3b68-4441-b776-2d79e25c3d0e"
      },
      "cell_type": "code",
      "source": [
        "vect.get_feature_names()[6489]"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'am'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "metadata": {
        "id": "DmagKWsjrKLk",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# CountVectorizer"
      ]
    },
    {
      "metadata": {
        "id": "bNkduzD2rKLl",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "This technique just consider the top max_features ordered by term frequency across the corpus."
      ]
    },
    {
      "metadata": {
        "id": "LCNgdhIzrKLl",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Using a CountVectorizer and a Random forest as classifier without any calibration we got an accuracy of 0.6322"
      ]
    },
    {
      "metadata": {
        "id": "3w0Yc_5_rKLm",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "ccf3ac7c-9ac6-4144-f244-1c0059eddbca"
      },
      "cell_type": "code",
      "source": [
        "# Default CountVectorizer\n",
        "vect = CountVectorizer()\n",
        "X_dtm = vect.fit_transform(X_train)\n",
        "\n",
        "# use Naive Bayes to predict the star rating\n",
        "np.random.seed(123)\n",
        "RF = RandomForestClassifier()\n",
        "pd.Series(cross_val_score(RF, X_dtm, y_train, cv=10)).describe()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "count    10.000000\n",
              "mean      0.632238\n",
              "std       0.017588\n",
              "min       0.591233\n",
              "25%       0.627551\n",
              "50%       0.635066\n",
              "75%       0.642966\n",
              "max       0.655454\n",
              "dtype: float64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "metadata": {
        "id": "cC_LrM-CrKLo",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "# define a function that accepts a vectorizer and calculates the accuracy\n",
        "def tokenize_test(vect):\n",
        "    X_dtm = vect.fit_transform(X_train)\n",
        "    print('Features: ', X_dtm.shape[1])\n",
        "    RF = RandomForestClassifier()\n",
        "    print(pd.Series(cross_val_score(RF, X_dtm, y_train, cv=10)).describe())"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "G8XnoeCArKLq",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "ef0325a2-89b0-42be-cffd-131f5b4b67f1"
      },
      "cell_type": "code",
      "source": [
        "# include 1-grams and 2-grams\n",
        "np.random.seed(123)\n",
        "vect = CountVectorizer(ngram_range=(1, 2))\n",
        "tokenize_test(vect)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Features:  87608\n",
            "count    10.000000\n",
            "mean      0.649163\n",
            "std       0.017322\n",
            "min       0.607543\n",
            "25%       0.643968\n",
            "50%       0.649337\n",
            "75%       0.661060\n",
            "max       0.667686\n",
            "dtype: float64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "fVswiqEirKLu",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## StopWords function"
      ]
    },
    {
      "metadata": {
        "id": "gapvErkIrKLw",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Using the Stopwords function we got an accuracy of 0.634376"
      ]
    },
    {
      "metadata": {
        "id": "Qc4a6XYgrKLw",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "6b253d10-f0ef-4e82-9d54-3c4c8d70deb1"
      },
      "cell_type": "code",
      "source": [
        "# remove English stop words\n",
        "np.random.seed(123)\n",
        "vect_1 = CountVectorizer(stop_words='english')\n",
        "tokenize_test(vect_1)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Features:  11777\n",
            "count    10.000000\n",
            "mean      0.634376\n",
            "std       0.019255\n",
            "min       0.603466\n",
            "25%       0.625860\n",
            "50%       0.638124\n",
            "75%       0.647044\n",
            "max       0.657492\n",
            "dtype: float64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "SytXqvkVrKLy",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Including ngrams"
      ]
    },
    {
      "metadata": {
        "id": "FGJSqsLKrKLy",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "23b8a474-e11a-444b-af70-630c183b0fa9"
      },
      "cell_type": "code",
      "source": [
        "# include 1-grams and 2-grams, and limit the number of features\n",
        "np.random.seed(123)\n",
        "vect_2 = CountVectorizer(ngram_range=(1, 2), max_features=1000)\n",
        "tokenize_test(vect_2)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Features:  1000\n",
            "count    10.000000\n",
            "mean      0.641109\n",
            "std       0.016062\n",
            "min       0.604485\n",
            "25%       0.637337\n",
            "50%       0.642529\n",
            "75%       0.646279\n",
            "max       0.661570\n",
            "dtype: float64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "YQIZf5lirKL2",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "We will include a \"cut off\" in the model using the min_df parameter, so this means that the model will ignore terms that have a document frequency strictly higher than the given threshold"
      ]
    },
    {
      "metadata": {
        "id": "Ij0Pvo3qrKL3",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "0d8cba12-08e3-44e8-b6e3-7334c0eaeeb9"
      },
      "cell_type": "code",
      "source": [
        "# include 1-grams and 2-grams, and only include terms that appear at least 2 times\n",
        "np.random.seed(123)\n",
        "vect = CountVectorizer(ngram_range=(1, 2), min_df=2)\n",
        "tokenize_test(vect)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Features:  22011\n",
            "count    10.000000\n",
            "mean      0.649570\n",
            "std       0.014517\n",
            "min       0.623853\n",
            "25%       0.644495\n",
            "50%       0.653238\n",
            "75%       0.655454\n",
            "max       0.675841\n",
            "dtype: float64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "4USwX3gbrKL5",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "#TfidfVectorizer"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "VWN3R-eYrKL6",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Stemming"
      ]
    },
    {
      "metadata": {
        "id": "ZMIV8_DgrKL8",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "stemmer = SnowballStemmer('english')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "gZrCxHuprKL-",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "# define a function that accepts text and returns a list of lemmas\n",
        "def split_into_steamer(text):\n",
        "    text = text.lower()\n",
        "    words = text.split()\n",
        "    return [stemmer.stem(word) for word in words]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "o_y8LwgJrKMA",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "aad557e4-26b5-4045-b2e1-500d9426666c"
      },
      "cell_type": "code",
      "source": [
        "np.random.seed(123)\n",
        "vect = CountVectorizer(analyzer=split_into_steamer)\n",
        "tokenize_test(vect)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Features:  18559\n",
            "count    10.000000\n",
            "mean      0.621226\n",
            "std       0.015219\n",
            "min       0.589195\n",
            "25%       0.621050\n",
            "50%       0.622834\n",
            "75%       0.630640\n",
            "max       0.638124\n",
            "dtype: float64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "En-eL-6PrKMD",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Lemmatization"
      ]
    },
    {
      "metadata": {
        "id": "REUK2iyyrKME",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "from nltk.stem import WordNetLemmatizer\n",
        "wordnet_lemmatizer = WordNetLemmatizer()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "ryjL0W4PrKMG",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "b2caf479-334b-4588-eb6f-19d5cac7327f"
      },
      "cell_type": "code",
      "source": [
        "import nltk\n",
        "nltk.download('wordnet')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[nltk_data] Downloading package wordnet to /Users/jfrank/nltk_data...\n",
            "[nltk_data]   Package wordnet is already up-to-date!\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "metadata": {
        "id": "0Q3rrqKlrKMK",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "# define a function that accepts text and returns a list of lemmas\n",
        "def split_into_lemmas(text):\n",
        "    text = text.lower()\n",
        "    words = text.split()\n",
        "    return [wordnet_lemmatizer.lemmatize(word) for word in words]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "_sOlHerCrKML",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "255fdfdf-033e-4e14-b384-7d22ec561c75"
      },
      "cell_type": "code",
      "source": [
        "# use split_into_lemmas as the feature extraction function (WARNING: SLOW!)\n",
        "np.random.seed(123)\n",
        "vect = CountVectorizer(analyzer=split_into_lemmas)\n",
        "tokenize_test(vect)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Features:  20084\n",
            "count    10.000000\n",
            "mean      0.620920\n",
            "std       0.019799\n",
            "min       0.583078\n",
            "25%       0.609993\n",
            "50%       0.618049\n",
            "75%       0.635066\n",
            "max       0.650357\n",
            "dtype: float64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "HSuAGRqmrKMN",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Finally we will prove using Tfidtfvectorizer to see which are the most common words on all this tweets"
      ]
    },
    {
      "metadata": {
        "id": "pSsJ5lPyrKMO",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "ec03e347-6c71-4365-f799-8a2bf0a1d303"
      },
      "cell_type": "code",
      "source": [
        "# create a document-term matrix using TF-IDF\n",
        "vect = TfidfVectorizer(stop_words='english')\n",
        "dtm = vect.fit_transform(X)\n",
        "features = vect.get_feature_names()\n",
        "dtm.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(14640, 14770)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "metadata": {
        "id": "12H7k9DgrKMQ",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "# create a dictionary of words and their TF-IDF scores\n",
        "word_scores = {}\n",
        "for word in vect.vocabulary_.keys():\n",
        "    word = word.lower()\n",
        "    if word in features:\n",
        "        word_scores[word] = dtm[review_id, features.index(word)]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "5F52cUXyrKMU",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "2824034a-d693-45a1-83bc-d3e2ea34c07f"
      },
      "cell_type": "code",
      "source": [
        "# print words with the top 5 TF-IDF scores\n",
        "print('TOP SCORING WORDS:')\n",
        "top_scores = sorted(word_scores.items(), key=lambda x: x[1], reverse=True)[:10]\n",
        "for word, score in top_scores:\n",
        "    print(word)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "TOP SCORING WORDS:\n",
            "hollywood\n",
            "dw5nf0ibtr\n",
            "los\n",
            "angeles\n",
            "mountains\n",
            "downtown\n",
            "rain\n",
            "view\n",
            "sign\n",
            "virginamerica\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "XgcR5CoZrKMW",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "We listed here the 10 most common words on our database and we found that the most common word is Hollywood follow it by and strange word, maybe is an user which has been mentioned a lot on this network."
      ]
    },
    {
      "metadata": {
        "id": "AdY5bnuNrKMX",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "2450f298-1990-4afd-a06e-5fb46be75045"
      },
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.figure(figsize=(12,6))\n",
        "plt.bar([\"CountVectorizer\",\"StopWords\",\"ngrams\",\"Stemming\", \"Lemmatization\"],[0.649,0.6343,0.649,0.6212, 0.6209],color=\"orange\",align='center')\n",
        "plt.title(\"Comparison of kernel Initializer\")\n",
        "plt.xlabel(\"\",fontsize=12)\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtQAAAF1CAYAAADIhX0mAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzt3XmcZGV97/HPlx1lNUxc2AYRF9zD\nCIhEMaKBaMC4AcEgXgyXRFziFrwagkS9inGLYhCjBiPI5pLBOwrKooGAYVAEhkURwRlRGRAQXICB\n3/3jnIZD2zNTPafP9PTM5/161avP8tRTv6pTXf3tp546lapCkiRJ0opZa7oLkCRJkmYyA7UkSZLU\ng4FakiRJ6sFALUmSJPVgoJYkSZJ6MFBLkiRJPRioJWkSkhyY5KzprmNMkg2TnJHk9iSnTbD/qCSf\nn47axtUxO0klWWcK+vpaklctY/9xSf5hxL7OS/KadnmVOraSZg4DtaRpkeQvk8xPcmeSn7Uhaffp\nrmt5qurEqnrBdNfR8TLg4cAfVNXLp7uYFZXk+iR7jtK2qvauqhPa6x2c5Pxx+w+rqn+abA2r4LGV\nNEMYqCWtdEneBHwEeC9NGNwG+ASw73TWtTxTMbo6gG2BH1TVkiFvZBW976s8HzdpzWCglrRSJdkU\nOBp4bVV9qap+XVX3VNUZVfXWts36ST6S5Mb28pEk67f79kiyKMnbktzUjm6/OMmfJflBkl8m+T+d\n2zsqyelJTklyR5LvJnlqZ/8RSX7U7rsyyV909h2c5IIkH07yS+Co7ohoGh9u67g9yWVJnjR2P5N8\nLsniJDckeWeStTr9np/kn5PcmuTHSfZexmP2hHZqwm1JFiTZp93+LuBIYL92pP+Q5Tz26yb5QpIv\nJlkvyVqd+39LklOTPKxtOzZF45AkPwHO6Wx7VZKfJLk5yTs6/S+1v8lY3uMzNk0jyROA44Bntvf/\ntnb/vyd5d7u8eZKvtsfh1nZ5q2Xdbrv8trbPscs9Sf693bdpkk+3z72fJnl3krU7fTzoOTPZ+y9p\n5jFQS1rZnglsAHx5GW3eAewKPA14KrAz8M7O/ke0fWxJEyg/BbwS2An4Y+DIJI/utN8XOA14GHAS\n8JUk67b7ftReZ1PgXcDnkzyyc91dgOuAPwTeM67OFwDPBh4LbAbsB9zS7vtY2+ejgecABwGvHtfv\nNcAWwDHAp5Nk/APR1nkGcFZbw+uAE5M8rqr+kWaU/5Sq2qiqPj3++p1+NgS+AtwFvKKq7gZeD7y4\nre9RwK3AseOu+hzgCcCfdrbtDjwOeB7NY/2Edvso/Y1quY9PVV0FHAZc2N7/zSboZy3gszQj+dsA\nvwU+vrwbr6pj2j43orn/i4FT290nAEuAxwBPp3kevGZc7Ut7zkhaDRmoJa1sfwDcvJwpCgcCR1fV\nTVW1mCbo/lVn/z3Ae6rqHuBkmtD10aq6o6oWAAuAp3TaX1JVp7ftP0QTxncFqKrTqurGqrqvqk4B\nfkgT4MfcWFUfq6olVfXbcXXeA2wMPB5IVV1VVT9rRyv3A97e1nQ98MFx9+GGqvpUVd1LE9AeSTP9\nZbxdgY2A91XV3VV1DvBV4IBlPH7jbQJ8neafh1e3twnwv4F3VNWiqrqLZjT1ZXnwNIWj2ncRuvf9\nXVX126r6PvB9mn96Ru1vVKM+PstUVbdU1Rer6jdVdQdNwH3OqNfv/CPy0aqal+ThwN7AG9vH5Sbg\nw8D+nast6zkjaTXk3C5JK9stwBZJ1llGqH4UcENn/YZ22/19dELhWGD5RWf/b2lC6JiFYwtVdV+S\nRWP9JTkIeBMwu22yEU1A/73rjldV5yT5OM0o7DZJvgy8BdgQWG+C+7BlZ/3nnX5+0w6+dmse8yhg\nYVXdt4y+lmdXYF3ggKqqzvZtgS8n6fZ9Lw8OrhPd/593ln/TqXuU/kY16uOzTEkeQhN49wI2bzdv\nnGTtznNoWT4NXFNV72/Xt6V5LH/WGTBfiwc/Tkt9zkhaPTlCLWlluxD4Hc3UgKW5kSa4jNmm3bai\nth5baOcxbwXcmGRbmukih9OcJWMz4AqgO7WgG0B/T1X9S1XtBDyRZurHW4GbaUavx9+Hn65A7TcC\nW4/Nv17Bvs4C/i9wdjvCOmYhsHdVbda5bFBV3b6Xef/HGaW/qba8+t5MMz1ll6rahGaKDjz4GE8o\nyRHtdbtz0xfSTJvZonMfN6mqJ06iJkmrGQO1pJWqqm6nmfd8bJoPEz6k/bDc3kmOaZt9AXhnkllJ\ntmjb9zmX8k5JXtJOPXgjTSC6CHgoTfhZDJDk1cCTRu00yTOS7NLOc/41zT8K97Yjn6cC70mycRvc\n37SC9+E7bd9vax+nPYA/p5nqMrKqOoZm/vjZ7WMKzQf63tPWR/t49znTylT3N4pfAFslWW8p+zem\necfitvYDkv84SqfthyBfD7y4O22jqn5G8w/KB5Ns0n4Qc/skI08jkbT6MVBLWumq6kM0AfOdNGF2\nIc0o8VfaJu8G5gOXAZcD3223raj/pJnTfCvNPOaXtGcWuZJmbvOFNMHsycAFk+h3E5oR7ltppmHc\nAvxzu+91NEH4OuB8mjD7mckW3n54cB+aebs305xe8KCqunoF+vonmsf4m224/CgwFzgryR00/2Ts\nMtl+O6a6v1GcQzNn/udJbp5g/0dopuDc3Nbz9RH73Q+YBVzVOdPHce2+g2im9FxJc+xPp5njLWkN\nlQdPp5Ok1UuSo4DHVNUrp7sWSdLqyRFqSZIkqQcDtSRJktSDUz4kSZKkHhyhliRJknowUEuSJEk9\nzLhvStxiiy1q9uzZ012GJEmSVnOXXHLJzVU1a3ntZlygnj17NvPnz5/uMiRJkrSaS3LDKO2c8iFJ\nkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqSJEnqwUAtSZIk\n9WCgliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg/rTHcBM8pJme4KNN5f1nRXIElanfi3ftUzA/7W\nO0ItSZIk9WCgliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHT5snSdJ4njpt1TMDTp2m\nNZcj1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerBQC1JkiT14Fk+JGlpPNPDqsmzPUhaxRiopeUx\nVK2aDFWSpFWEUz4kSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4FakiRJ6mHQQJ1kryTXJLk2yRFL\nafOKJFcmWZDkpCHrkSRJkqbaYKfNS7I2cCzwfGARcHGSuVV1ZafNDsDbgWdV1a1J/nCoeiRJkqQh\nDDlCvTNwbVVdV1V3AycD+45r89fAsVV1K0BV3TRgPZIkSdKUGzJQbwks7Kwvard1PRZ4bJILklyU\nZK8B65EkSZKm3JDflDjR18uN/2qzdYAdgD2ArYD/SvKkqrrtQR0lhwKHAmyzzTZTX6kkSZK0goYc\noV4EbN1Z3wq4cYI2/1lV91TVj4FraAL2g1TV8VU1p6rmzJo1a7CCJUmSpMkaMlBfDOyQZLsk6wH7\nA3PHtfkK8FyAJFvQTAG5bsCaJEmSpCk1WKCuqiXA4cCZwFXAqVW1IMnRSfZpm50J3JLkSuBc4K1V\ndctQNUmSJElTbcg51FTVPGDeuG1HdpYLeFN7kSRJkmYcvylRkiRJ6sFALUmSJPVgoJYkSZJ6MFBL\nkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerBQC1JkiT1YKCWJEmSejBQS5Ik\nST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqSJEnqwUAtSZIk9WCgliRJknowUEuSJEk9\nGKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4FakiRJ6sFALUmSJPVgoJYkSZJ6MFBLkiRJPRio\nJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerBQC1JkiT1YKCWJEmSejBQS5IkST0YqCVJ\nkqQeDNSSJElSDwZqSZIkqYdBA3WSvZJck+TaJEdMsP/gJIuTXNpeXjNkPZIkSdJUW2eojpOsDRwL\nPB9YBFycZG5VXTmu6SlVdfhQdUiSJElDGnKEemfg2qq6rqruBk4G9h3w9iRJkqSVbshAvSWwsLO+\nqN023kuTXJbk9CRbT9RRkkOTzE8yf/HixUPUKkmSJK2QIQN1JthW49bPAGZX1VOAbwInTNRRVR1f\nVXOqas6sWbOmuExJkiRpxQ0ZqBcB3RHnrYAbuw2q6paquqtd/RSw04D1SJIkSVNuyEB9MbBDku2S\nrAfsD8ztNkjyyM7qPsBVA9YjSZIkTbnBzvJRVUuSHA6cCawNfKaqFiQ5GphfVXOB1yfZB1gC/BI4\neKh6JEmSpCEMFqgBqmoeMG/ctiM7y28H3j5kDZIkSdKQ/KZESZIkqQcDtSRJktSDgVqSJEnqwUAt\nSZIk9WCgliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4FakiRJ6sFALUmS\nJPVgoJYkSZJ6MFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerBQC1JkiT1\nYKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqSJEnqwUAtSZIk9WCg\nliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4FakiRJ6sFALUmSJPVgoJYk\nSZJ6MFBLkiRJPRioJUmSpB4GDdRJ9kpyTZJrkxyxjHYvS1JJ5gxZjyRJkjTVBgvUSdYGjgX2BnYE\nDkiy4wTtNgZeD3xnqFokSZKkoQw5Qr0zcG1VXVdVdwMnA/tO0O6fgGOA3w1YiyRJkjSIIQP1lsDC\nzvqidtv9kjwd2LqqvrqsjpIcmmR+kvmLFy+e+kolSZKkFTRkoM4E2+r+nclawIeBNy+vo6o6vqrm\nVNWcWbNmTWGJkiRJUj9DBupFwNad9a2AGzvrGwNPAs5Lcj2wKzDXDyZKkiRpJhkyUF8M7JBkuyTr\nAfsDc8d2VtXtVbVFVc2uqtnARcA+VTV/wJokSZKkKTVYoK6qJcDhwJnAVcCpVbUgydFJ9hnqdiVJ\nkqSVaZ0hO6+qecC8cduOXErbPYasRZIkSRqC35QoSZIk9WCgliRJknpYbqBOcniSzVdGMZIkSdJM\nM8oI9SOAi5OcmmSvJBOdX1qSJElaIy03UFfVO4EdgE8DBwM/TPLeJNsPXJskSZK0yhtpDnVVFfDz\n9rIE2Bw4PckxA9YmSZIkrfKWe9q8JK8HXgXcDPwb8Naquqf96vAfAm8btkRJkiRp1TXKeai3AF5S\nVTd0N1bVfUleNExZkiRJ0swwypSPecAvx1aSbJxkF4CqumqowiRJkqSZYJRA/a/AnZ31X7fbJEmS\npDXeKIE67YcSgWaqBwN/ZbkkSZI0U4wSqK9L8vok67aXNwDXDV2YJEmSNBOMEqgPA3YDfgosAnYB\nDh2yKEmSJGmmWO7Ujaq6Cdh/JdQiSZIkzTijnId6A+AQ4InABmPbq+p/DViXJEmSNCOMMuXjP4BH\nAH8KfAvYCrhjyKIkSZKkmWKUQP2YqvoH4NdVdQLwQuDJw5YlSZIkzQyjBOp72p+3JXkSsCkwe7CK\nJEmSpBlklPNJH59kc+CdwFxgI+AfBq1KkiRJmiGWGaiTrAX8qqpuBb4NPHqlVCVJkiTNEMuc8tF+\nK+LhK6kWSZIkacYZZQ71N5K8JcnWSR42dhm8MkmSJGkGGGUO9dj5pl/b2VY4/UOSJEka6ZsSt1sZ\nhUiSJEkz0SjflHjQRNur6nNTX44kSZI0s4wy5eMZneUNgOcB3wUM1JIkSVrjjTLl43Xd9SSb0nwd\nuSRJkrTGG+UsH+P9BthhqguRJEmSZqJR5lCfQXNWD2gC+I7AqUMWJUmSJM0Uo8yh/ufO8hLghqpa\nNFA9kiRJ0owySqD+CfCzqvodQJINk8yuqusHrUySJEmaAUaZQ30acF9n/d52myRJkrTGGyVQr1NV\nd4+ttMvrDVeSJEmSNHOMEqgXJ9lnbCXJvsDNw5UkSZIkzRyjzKE+DDgxycfb9UXAhN+eKEmSJK1p\nRvlilx8BuybZCEhV3TF8WZIkSdLMsNwpH0nem2Szqrqzqu5IsnmSd6+M4iRJkqRV3ShzqPeuqtvG\nVqrqVuDPhitJkiRJmjlGCdRrJ1l/bCXJhsD6y2gvSZIkrTFGCdSfB85OckiSQ4BvACeM0nmSvZJc\nk+TaJEdMsP+wJJcnuTTJ+Ul2nFz5kiRJ0vQa5UOJxyS5DNgTCPB1YNvlXS/J2sCxwPNpzgxycZK5\nVXVlp9lJVXVc234f4EPAXpO+F5IkSdI0GWWEGuDnNN+W+FLgecBVI1xnZ+Daqrqu/TKYk4F9uw2q\n6led1YcCNWI9kiRJ0iphqSPUSR4L7A8cANwCnEJz2rznjtj3lsDCzvoiYJcJbue1wJtovn3xT5ZS\ny6HAoQDbbLPNiDcvSZIkDW9ZI9RX04xG/3lV7V5VHwPunUTfmWDb741AV9WxVbU98PfAOyfqqKqO\nr6o5VTVn1qxZkyhBkiRJGtayAvVLaaZ6nJvkU0mex8QheWkWAVt31rcCblxG+5OBF0+if0mSJGna\nLTVQV9WXq2o/4PHAecDfAQ9P8q9JXjBC3xcDOyTZLsl6NNNH5nYbJNmhs/pC4IeTrF+SJEmaVsv9\nUGJV/bqqTqyqF9GMMl8K/N4p8Ca43hLgcOBMmg8xnlpVC5Ic3Z7RA+DwJAuSXEozj/pVK3pHJEmS\npOmw3NPmdVXVL4FPtpdR2s8D5o3bdmRn+Q2TuX1JkiRpVTPqafMkSZIkTcBALUmSJPVgoJYkSZJ6\nMFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerBQC1JkiT1YKCWJEmSejBQ\nS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqSJEnqwUAtSZIk9WCgliRJknowUEuS\nJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4FakiRJ6sFALUmSJPVgoJYkSZJ6MFBLkiRJ\nPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerBQC1JkiT1YKCWJEmSejBQS5IkST0Y\nqCVJkqQeDNSSJElSDwZqSZIkqYdBA3WSvZJck+TaJEdMsP9NSa5MclmSs5NsO2Q9kiRJ0lQbLFAn\nWRs4Ftgb2BE4IMmO45p9D5hTVU8BTgeOGaoeSZIkaQhDjlDvDFxbVddV1d3AycC+3QZVdW5V/aZd\nvQjYasB6JEmSpCk3ZKDeEljYWV/UbluaQ4CvTbQjyaFJ5ieZv3jx4iksUZIkSepnyECdCbbVhA2T\nVwJzgA9MtL+qjq+qOVU1Z9asWVNYoiRJktTPOgP2vQjYurO+FXDj+EZJ9gTeATynqu4asB5JkiRp\nyg05Qn0xsEOS7ZKsB+wPzO02SPJ04JPAPlV104C1SJIkSYMYLFBX1RLgcOBM4Crg1KpakOToJPu0\nzT4AbAScluTSJHOX0p0kSZK0ShpyygdVNQ+YN27bkZ3lPYe8fUmSJGloflOiJEmS1IOBWpIkSerB\nQC1JkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqSJEnqwUAt\nSZIk9WCgliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4FakiRJ6sFALUmS\nJPVgoJYkSZJ6MFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerBQC1JkiT1\nYKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqSJEnqwUAtSZIk9WCg\nliRJknowUEuSJEk9GKglSZKkHgYN1En2SnJNkmuTHDHB/mcn+W6SJUleNmQtkiRJ0hAGC9RJ1gaO\nBfYGdgQOSLLjuGY/AQ4GThqqDkmSJGlI6wzY987AtVV1HUCSk4F9gSvHGlTV9e2++wasQ5IkSRrM\nkFM+tgQWdtYXtdskSZKk1caQgToTbKsV6ig5NMn8JPMXL17csyxJkiRp6gwZqBcBW3fWtwJuXJGO\nqur4qppTVXNmzZo1JcVJkiRJU2HIQH0xsEOS7ZKsB+wPzB3w9iRJkqSVbrBAXVVLgMOBM4GrgFOr\nakGSo5PsA5DkGUkWAS8HPplkwVD1SJIkSUMY8iwfVNU8YN64bUd2li+mmQoiSZIkzUh+U6IkSZLU\ng4FakiRJ6sFALUmSJPVgoJYkSZJ6MFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOB\nWpIkSerBQC1JkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqS\nJEnqwUAtSZIk9WCgliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4FakiRJ\n6sFALUmSJPVgoJYkSZJ6MFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerB\nQC1JkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSD4MG6iR7JbkmybVJjphg//pJTmn3fyfJ\n7CHrkSRJkqbaYIE6ydrAscDewI7AAUl2HNfsEODWqnoM8GHg/UPVI0mSJA1hyBHqnYFrq+q6qrob\nOBnYd1ybfYET2uXTgeclyYA1SZIkSVNqyEC9JbCws76o3TZhm6paAtwO/MGANUmSJElTap0B+55o\npLlWoA1JDgUObVfvTHJNz9rWdFsAN093EVPiQN/QmCSP/ZrLY7/mWj2Ovcd9RXjs+9t2lEZDBupF\nwNad9a2AG5fSZlGSdYBNgV+O76iqjgeOH6jONU6S+VU1Z7rr0MrnsV9zeezXXB77NZfHfuUZcsrH\nxcAOSbZLsh6wPzB3XJu5wKva5ZcB51TV741QS5IkSauqwUaoq2pJksOBM4G1gc9U1YIkRwPzq2ou\n8GngP5JcSzMyvf9Q9UiSJElDGHLKB1U1D5g3btuRneXfAS8fsgZNyOkzay6P/ZrLY7/m8tivuTz2\nK0mcYSFJkiStOL96XJIkSerBQD2gJI9IcnKSHyW5Msm8JI+dwv73SLJbZ/nCcfvXSfKLJI+cZL+b\nJfnbFazpv1fkequrJHdOdw1d449tkkclOX26+tEDkrwjyYIklyW5NMkuSd6Y5CEruY7Dkhy0Mm9z\nTTXVxzzJU5Nc2lk/IMlvkqzbrj85yWU96p2d5IoVvb60OjNQD6T9xscvA+dV1fZVtSPwf4CHT+HN\n7AHs1i5/G9gqyezO/j2BK6rqZ5PsdzNgUoG6/ap5qmq35bUdtS8N4kHHtqpurKqXTWM/ApI8E3gR\n8EdV9RSa392FwBuBlRqoq+q4qvrcyrzNNdFAx/xyYNskG7fruwFXA0/vrF8wiRp9LZ5h2lMQaxoY\nqIfzXOCeqjpubENVXQqcn+QDSa5IcnmS/eD+EeavjrVN8vEkB7fL1yd5V5Lvttd5fBucDwP+rh2R\neBZwGrBfp4b9gS+0fWyf5OtJLknyX0ke325/eJIvJ/l+e9kNeB+wfTti8oE0llbzuUlOonkhv39E\nNsnR7fUvTfLTJJ9tt78yyf+02z859oKd5M72Ot8BnjmVB2JVk2RWki8mubi9PKvdflSSE5Kc1R7z\nlyQ5pn3Mv94ZZbo+yXuTXJhkfpI/SnJmmndCDmvbbJTk7M5zZt/25scf2/tHnJL8W+eYLU7yjyvY\nzwZJPtu2/16S57bbD07ypfa+/DDJMSvvUV/lPRK4uaruAqiqm2lOJfoo4Nwk5wIkeUF73L+b5LQk\nG7XbR3lO7JHkW0lOTfKDJO9LcmD7+3h5ku3bdkcleUu7fF6S97dtfpDkj9vtD2n7uSzJKUm+k8Rz\n3U7OqMf8gPb4XJHk/WNXbl8zP9g+F85OMquq7qM5Ze0ubbOdgGN5YOBlN+C/2+s/r/39vDzJZ5Ks\n326/PsmRSc4HXp5kpzR/Gy4EXtu5/Sd2XssvS7LDYI/UGqZ9Pb0qyafSvINxVpINkzyjfawvbF93\nx15zD25fD84Azlra63bb79Xta/0VSU5MsmeSC9rX5J3bds/p/C34Xh74B03LUlVeBrgArwc+PMH2\nlwLfoDmV4MOBn9C8sO4BfLXT7uPAwe3y9cDr2uW/Bf6tXT4KeEvnOs8Avtcurw/cBGzerp8N7NAu\n70Jzzm+AU4A3tstr03y5zmyake1Rav41sF2n7Z3j7u+mwGU0L+xPAM4A1m33fQI4qF0u4BXTfdwG\neB7cOcG2k4Dd2+VtgKs6x/N8YF3gqcBvgL3bfV8GXtx5PvxNu/zh9vHdGJgF3NRuXwfYpF3eAriW\n5ptJxx/bB62327alGdXadkX6Ad4MfLZdfnz7fNkAOBi4rn1ObADcAGw93cdoVbgAGwGXAj9ofy+e\n0znWW3Qe/28DD23X/x44chLPiT2A29rf3fWBnwLvave9AfhI53n4lnb5POCD7fKfAd9sl98CfLJd\nfhKwBJgz3Y/jTLqMeMwf1f7+zGp/F8/pvA4UcGC7fCTw8c7xOxJ4KM3ryfbAqe2+HwKPbn//FgKP\nbbd/jgf+DlwPvK1T52Wd2j7Q+T3/WOf21wM2nO7HdHW50LyeLgGe1q6fCrwSuALYrd32vs6xOJjm\ni/Ie1q4v63V7CfBkmgHVS4DPtPv2Bb7SXucM4Fmd5+k60/2YzISLbw2sfLsDX6iqe4FfJPkWTRD+\n1XKu96X25yXASyZqUFUXt/+ZPo4mvF5UVbe2o1i7Aacl93995/rtzz8BDmqvfy9we5LNJ1Hz/1TV\njyeqJ82NnUjzj8Ulac5LvhNwcVvHhjShH+Be4IvLeQxWF3sCO3aOxSadEYCvVdU9SS6n+Qfm6+32\ny2leDMfM7WzfqKruAO5I8rskm9H8o/PeJM8G7gO2ZITpRkk2oHmn4/CquiHNqPhk+9md5o8tVXV1\nkhuAsc8OnF1Vt7e3dSVNaF+OP/ddAAAFH0lEQVS4vLpWd1V1Z5KdgD+meXfrlCRHjGu2K7AjcEH7\n3FkP6H5uYnnPCYCLq50CluRHwFmd6zx3KeV1X3tmt8u7Ax9ta78iPeblrqlGPObPoJk2uBggyYnA\ns4Gv0Pw+ntK2+zwPHKcLaP6p/S+a4/2jJI9JMovmeXFdkqcCP66qH7TXOYFm9Pkj7fop7e1tCmxW\nVd9qt/8HsHe7fCHwjiRbAV+qqh/2fEj0YD+u5l1teOB3b+OqGvuc0kk0U4bGfKOqxr5pOiz9dfvH\nVTX2jvICmtfkav/mzG7bXAB8qH2+famqFk35vVsNGaiHs4Dm7bvxlvaF9Et48BScDcbtv6v9eS/L\nPm4n00z1eALtdI+239uq6mnLKngZllYzNMFtaY4CFlXVZzv9nFBVb5+g7e/awL4mWAt4ZlX9trux\nDUljb//el+SeaocIaF4Uu8f9rs72uzrbx9odSDOqtVMb0K/n959TEzmO5gX0m+36ivSzrOdLt9bl\nPZfXKO3z/zzgvPaP26vGNQnNH80DltLF8p4TTLD9rgnaLK3f7vFa1jHWiEY85iN31/68iCaI784D\n/3Atovm7MBbGltfv2Ot6Ov0++MaqTkozRe+FwJlJXlNV50yiXi3b+NfKRy2nffdv8bJet5f7GlBV\n70vy/2jelbooyZ5VdfUK3Ys1iHOoh3MOsH6Svx7bkOQZwK3AfknWbkcMng38D83b3zsmWb8dFXje\nCLdxB83bul1foHlr6E9oR6yq6lfAj5O8vK0j7QgFNFNB/qbdvnaSTSbo99tLqXmpkrwIeD7N1Jcx\nZwMvS/KHbZuHJdl2hPu5ujkLOHxsJcmK/qOzLJvSvNV/T5o5zGOP80TPmbE6XkszAvK+Pv3QPF8O\nbPt8LM20lmv63JnVXZLHjZuD+jSa14Tu43wR8Kwkj2mv85BM4VmDJul84BVtHTvSvIWsSRjxmH8H\neE6SLdJ83uQAYGy0eC0eGLT5S5pjQvvOxEKaaQBjgfpCmg87jgXqq4HZY88l4K86/d6vqm6jeddy\n93bTgZ36Hw1cV1X/QvO35imTuf+atFtp3nHatV1f1jdLL+11eyRJtq+qy6vq/cB8mql7Wg4D9UDa\nkcW/AJ6f5oNBC2hGbE+imZP2fZrQ/baq+nlVLaSZJ3UZzTSJ741wM2cAf9F+cOCP29u9kmbu7TlV\nNf4/1kOSfJ9m9Hzsw2VvAJ7bjo5cAjyxqm6heVv5iiQfoJm/+3s1L6e2N9P8Rz32oZWj29reSfOh\nicto5mVP6pR+M9BDkizqXN5E80/GnDQfLrmS5sOlU+3E9jbm0xz7qwEmOLZdbwGe3PkwymEr2M8n\ngLXb59QpNJ8FuAsty0bACWlOr3kZzdSOo2i+5exrSc5t3/Y/GPhC2+Yipu8P3SeAWW0df0/z+nD7\nNNUyU41yzH8GvB04l+b197tV9Z/t9X8NPDHJJTQDKEd3+r4AWL/9uwJNoH40baCu5luKX00zDfBy\nmtHJ45jYq4Fj03wosfuu2n7AFWk+FP94mnnYGtYhwPHtsQhL/52b8HV7Et7YvrZ/n+aYf21FC16T\n+E2JkqRJaUdL162q36U5O8jZNB9wu3uaS1tjJLmzqjaa7jq08iTZqKrGzqR1BPDIqnrDNJellvMX\nJUmT9RCaU7utSzNS9jeGaWlwL0zydprsdgPNO1ZaRThCLUmSJPXgHGpJkiSpBwO1JEmS1IOBWpIk\nSerBQC1JkiT1YKCWJEmSejBQS5IkST38f2Ej9Sd4IsvEAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x1a31cdb668>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "H_4gteA7rKMZ",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "the best model was the ngrams one, so we will use it to train the Neural Network"
      ]
    },
    {
      "metadata": {
        "id": "PtrMGdxlrKMZ",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Exercise 8.2\n",
        "\n",
        "Train a Deep Neural Network with the following architecture:\n",
        "\n",
        "- Input = text \n",
        "- Dense(128)\n",
        "- Relu Activation\n",
        "- BatchNormalization\n",
        "- Dropout(0.5)\n",
        "- Dense(10, Softmax)\n",
        "\n",
        "Optimized using rmsprop using as loss categorical_crossentropy\n",
        "\n",
        "Hints: \n",
        "- test with two iterations then try more. \n",
        "- learning can be ajusted\n",
        "\n",
        "Evaluate the performance using the testing set (aprox 55% with 50 epochs)"
      ]
    },
    {
      "metadata": {
        "id": "CcsHxyburKMc",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "from keras.models import Sequential\n",
        "from keras.utils import np_utils\n",
        "from keras.layers import Dense, Dropout, Activation, BatchNormalization\n",
        "from keras.optimizers import rmsprop\n",
        "from keras.callbacks import History\n",
        "from livelossplot import PlotLossesKeras"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Nw6H19vfrKMd",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "1c92c927-4669-4926-a0bc-88822c91794a"
      },
      "cell_type": "code",
      "source": [
        "model = Sequential()\n",
        "model.add(Dense(128, input_shape=(22011,)))\n",
        "model.add(Activation('relu'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(Dropout(0.5,seed=123))\n",
        "model.add(Dense(3, activation='softmax'))\n",
        "model.summary()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_7 (Dense)              (None, 128)               2817536   \n",
            "_________________________________________________________________\n",
            "activation_4 (Activation)    (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "batch_normalization_4 (Batch (None, 128)               512       \n",
            "_________________________________________________________________\n",
            "dropout_4 (Dropout)          (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_8 (Dense)              (None, 3)                 387       \n",
            "=================================================================\n",
            "Total params: 2,818,435\n",
            "Trainable params: 2,818,179\n",
            "Non-trainable params: 256\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "TqxzrdjbrKMf",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=rmsprop(lr=0.0001), metrics=[\"accuracy\"])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "9U9-DogOrKMg",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "X_dtm = vect.fit_transform(X_train)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "h5dD-YRUrKMk",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "e155dc23-0ca8-4cf6-8bdc-c4b01e14674a"
      },
      "cell_type": "code",
      "source": [
        "X_dtm.shape, y_train.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((9808, 22011), (9808, 3))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 53
        }
      ]
    },
    {
      "metadata": {
        "id": "3HvWTn9yrKMn",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "6a991de6-0d70-4af2-e75e-cff03830149c"
      },
      "cell_type": "code",
      "source": [
        "np.random.seed(123)\n",
        "batch_size = 1000\n",
        "model.fit(X_dtm, y_train, epochs = 50, verbose = 2, batch_size=batch_size, validation_split=0.1,callbacks=[PlotLossesKeras()])\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAAE1CAYAAAD6akEFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x1a2e5c69b0>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x1a2f92d908>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        }
      ]
    },
    {
      "metadata": {
        "id": "LELb24LurKMp",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "X_test_1 = vect.transform(X_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "zEw5zSOKrKMq",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "model_pred_1 = model.predict_classes(X_test_1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "MKK80TdMrKMs",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        "y_test_1 = y_test.argmax(axis=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "G0TCUWaMrKMt",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "a71aee85-f147-4dbe-c786-9f7733f318c9"
      },
      "cell_type": "code",
      "source": [
        "print('the prediction model is=',(model_pred_1==y_test_1).mean())"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "the prediction model is= 0.792632450331\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "F1bncYr8rKMv",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "All the previous models got an accuracy between 0.62 and 0.65 and using a simple Neural Network the accuracy of the model increase between 0.78 and 0.8."
      ]
    },
    {
      "metadata": {
        "id": "jh5yCGiWrKMw",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        },
        "outputId": "02c1f24d-2ec8-4f32-9091-f725dbdea455"
      },
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.figure(figsize=(12,6))\n",
        "plt.bar([\"CountVectorizer\",\"StopWords\",\"ngrams\",\"Stemming\", \"Lemmatization\",\"NN\"],[0.649,0.6343,0.649,0.6212, 0.6209, 0.7926],color=\"orange\",align='center')\n",
        "plt.title(\"Comparison of kernel Initializer\")\n",
        "plt.xlabel(\"\",fontsize=12)\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtQAAAF1CAYAAADIhX0mAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzt3XucXWV97/HP13DTctMmtUqAIOIl\n3o8RFG3Fgi3YHrBWC6hH8dBy6BHxUmqxUoq09SjWS1uxilVLrcit1QZPFCpgWz1gExSBgGhEMBGt\nAcGCFyDwO3+sNboYZjI7s2ZlZpLP+/Wa16zLs5/923vt2fs7z3722qkqJEmSJE3Pg2a7AEmSJGk+\nM1BLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpI2QZKXJblotusYk+TBSS5I8oMk\n502w/5Qk/zAbtY2rY0mSSrLNDPT16SSv3Mj+9yf54xH7+lyS32mX59SxlTR/GKglzYokL02yKsmd\nSb7ThqTnzHZdU6mqj1XVr852HR0vBh4O/HxVvWS2i5muJDcmOWiUtlV1SFWd2V7uqCSfH7f/2Kr6\n002tYQ4eW0nzhIFa0maX5A3Ae4C30oTBPYD3AYfNZl1TmYnR1QHsCXytqjYMeSVz9LbPed5v0tbB\nQC1ps0qyC3Aq8Oqq+qeq+mFV3VNVF1TVH7Rttk/yniQ3tz/vSbJ9u++AJOuSvDHJ99rR7RcmeUGS\nryX5fpI/6lzfKUnOT3JOkjuSfCnJUzr7T0zyjXbftUl+s7PvqCRfSPLuJN8HTumOiKbx7raOHyS5\nKskTx25nkr9Psj7JTUlOSvKgTr+fT/IXSW5L8s0kh2zkPnt8OzXh9iSrkxzabn8LcDJweDvSf/QU\n9/22ST6e5B+TbJfkQZ3bf2uSc5M8rG07NkXj6CTfAi7pbHtlkm8luSXJmzv9T9rfppjq/hmbppHk\n8cD7gWe1t//2dv/fJfmzdvmhST7VHofb2uXFG7vedvmNbZ9jP/ck+bt23y5JPtQ+9r6d5M+SLOj0\ncb/HzKbefknzj4Fa0ub2LGAH4BMbafNm4JnAU4GnAPsCJ3X2/2Lbx240gfKDwMuBpwO/BJyc5FGd\n9ocB5wEPA84CPplk23bfN9rL7AK8BfiHJI/oXHY/4AbgF4A/H1fnrwK/DDwG2BU4HLi13ffXbZ+P\nAp4LvAJ41bh+rwcWAqcBH0qS8XdEW+cFwEVtDa8BPpbksVX1JzSj/OdU1Y5V9aHxl+/082Dgk8Bd\nwG9X1d3A8cAL2/oeCdwGnD7uos8FHg/8Wmfbc4DHAgfS3NePb7eP0t+oprx/quo64Fjgsvb27zpB\nPw8CPkIzkr8H8GPgvVNdeVWd1va5I83tXw+c2+4+E9gAPBp4Gs3j4HfG1T7ZY0bSFshALWlz+3ng\nlimmKLwMOLWqvldV62mC7v/o7L8H+POqugc4myZ0/WVV3VFVq4HVwJM77a+oqvPb9u+iCePPBKiq\n86rq5qq6r6rOAb5OE+DH3FxVf11VG6rqx+PqvAfYCXgckKq6rqq+045WHg68qa3pRuCd427DTVX1\nwaq6lyagPYJm+st4zwR2BN5WVXdX1SXAp4AjN3L/jbcz8Bmafx5e1V4nwP8C3lxV66rqLprR1Bfn\n/tMUTmnfReje9rdU1Y+r6ivAV2j+6Rm1v1GNev9sVFXdWlX/WFU/qqo7aALuc0e9fOcfkb+sqhVJ\nHg4cAryuvV++B7wbOKJzsY09ZiRtgZzbJWlzuxVYmGSbjYTqRwI3ddZvarf9tI9OKBwLLP/Z2f9j\nmhA6Zu3YQlXdl2TdWH9JXgG8AVjSNtmRJqA/4LLjVdUlSd5LMwq7R5JPACcADwa2m+A27NZZ/26n\nnx+1g6/dmsc8ElhbVfdtpK+pPBPYFjiyqqqzfU/gE0m6fd/L/YPrRLf/u53lH3XqHqW/UY16/2xU\nkofQBN6DgYe2m3dKsqDzGNqYDwHXV9Xb2/U9ae7L73QGzB/E/e+nSR8zkrZMjlBL2twuA35CMzVg\nMjfTBJcxe7Tbpmv3sYV2HvNi4OYke9JMFzmO5iwZuwLXAN2pBd0A+gBV9VdV9XTgCTRTP/4AuIVm\n9Hr8bfj2NGq/Gdh9bP71NPu6CPg/wMXtCOuYtcAhVbVr52eHqur2vdHbP84o/c20qer7fZrpKftV\n1c40U3Tg/sd4QklObC/bnZu+lmbazMLObdy5qp6wCTVJ2sIYqCVtVlX1A5p5z6en+TDhQ9oPyx2S\n5LS22ceBk5IsSrKwbd/nXMpPT/KidurB62gC0eXAz9GEn/UASV4FPHHUTpM8I8l+7TznH9L8o3Bv\nO/J5LvDnSXZqg/sbpnkbvtj2/cb2fjoA+O80U11GVlWn0cwfv7i9T6H5QN+ft/XR3t99zrQy0/2N\n4j+BxUm2m2T/TjTvWNzefkDyT0bptP0Q5PHAC7vTNqrqOzT/oLwzyc7tBzH3TjLyNBJJWx4DtaTN\nrqreRRMwT6IJs2tpRok/2Tb5M2AVcBVwNfCldtt0/TPNnObbaOYxv6g9s8i1NHObL6MJZk8CvrAJ\n/e5MM8J9G800jFuBv2j3vYYmCN8AfJ4mzH54UwtvPzx4KM283VtoTi/4iqr66jT6+lOa+/izbbj8\nS2A5cFGSO2j+ydhvU/vtmOn+RnEJzZz57ya5ZYL976GZgnNLW89nRuz3cGARcF3nTB/vb/e9gmZK\nz7U0x/58mjnekrZSuf90OknasiQ5BXh0Vb18tmuRJG2ZHKGWJEmSejBQS5IkST045UOSJEnqwRFq\nSZIkqQcDtSRJktTDvPumxIULF9aSJUtmuwxJkiRt4a644opbqmrRVO3mXaBesmQJq1atmu0yJEmS\ntIVLctMo7ZzyIUmSJPVgoJYkSZJ6MFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1MOg\ngTrJwUmuT7ImyYkT7N8jyaVJvpzkqiQvGLIeSZIkaaYNFqiTLABOBw4BlgJHJlk6rtlJwLlV9TTg\nCOB9Q9UjSZIkDWHIEep9gTVVdUNV3Q2cDRw2rk0BO7fLuwA3D1iPJEmSNOOGDNS7AWs76+vabV2n\nAC9Psg5YAbxmoo6SHJNkVZJV69evH6JWSZIkaVqGDNSZYFuNWz8S+LuqWgy8APhokgfUVFVnVNWy\nqlq2aNGiAUqVJEmSpmfIQL0O2L2zvpgHTuk4GjgXoKouA3YAFg5YkyRJkjSjthmw75XAPkn2Ar5N\n86HDl45r8y3gQODvkjyeJlA7p0PaUp010RtXmraXjn/TT5I0GwYboa6qDcBxwIXAdTRn81id5NQk\nh7bNfh/43SRfAT4OHFVVvkJIkiRp3hhyhJqqWkHzYcPutpM7y9cCzx6yBkmSJGlIflOiJEmS1IOB\nWpIkSerBQC1JkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqS\nJEnqwUAtSZIk9WCgliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4FakiRJ\n6sFALUmSJPVgoJYkSZJ6MFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerB\nQC1JkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeBg3USQ5Ocn2SNUlOnGD/u5Nc2f58LcntQ9YjSZIk\nzbRthuo4yQLgdOD5wDpgZZLlVXXtWJuqen2n/WuApw1VjyRJkjSEIUeo9wXWVNUNVXU3cDZw2Eba\nHwl8fMB6JEmSpBk3ZKDeDVjbWV/XbnuAJHsCewGXTLL/mCSrkqxav379jBcqSZIkTdeQgToTbKtJ\n2h4BnF9V9060s6rOqKplVbVs0aJFM1agJEmS1NeQgXodsHtnfTFw8yRtj8DpHpIkSZqHhgzUK4F9\nkuyVZDua0Lx8fKMkjwUeClw2YC2SJEnSIAYL1FW1ATgOuBC4Dji3qlYnOTXJoZ2mRwJnV9Vk00Ek\nSZKkOWuw0+YBVNUKYMW4bSePWz9lyBokSZKkIflNiZIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerB\nQC1JkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqSJEnqwUAt\nSZIk9WCgliRJknrYZrYLmFfOymxXsOV4ac12BZIkDcO8MLPmQWZwhFqSJEnqwUAtSZIk9WCgliRJ\nknowUEuSJEk9GKglSZKkHjzLhyRJ84lnkJhZ8+AMEpr7HKGWJEmSejBQS5IkST0YqCVJkqQeDNSS\nJElSD34oUVsOP6gzs/ygztbJv6OZ5d+RtFVwhFqSJEnqwUAtSZIk9WCgliRJknowUEuSJEk9DBqo\nkxyc5Poka5KcOEmb305ybZLVSc4ash5JkiRppg12lo8kC4DTgecD64CVSZZX1bWdNvsAbwKeXVW3\nJfmFoeqRJEmShjDkCPW+wJqquqGq7gbOBg4b1+Z3gdOr6jaAqvregPVIkiRJM27IQL0bsLazvq7d\n1vUY4DFJvpDk8iQHD1iPJEmSNOOG/GKXib4dYPwZ7rcB9gEOABYD/57kiVV1+/06So4BjgHYY489\nZr5SSZIkaZqGHKFeB+zeWV8M3DxBm3+uqnuq6pvA9TQB+36q6oyqWlZVyxYtWjRYwZIkSdKmGjJQ\nrwT2SbJXku2AI4Dl49p8EngeQJKFNFNAbhiwJkmSJGlGDRaoq2oDcBxwIXAdcG5VrU5yapJD22YX\nArcmuRa4FPiDqrp1qJokSZKkmTbkHGqqagWwYty2kzvLBbyh/ZEkSZLmHb8pUZIkSerBQC1JkiT1\nYKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqSJEnqwUAtSZIk9WCg\nliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4FakiRJ6sFALUmSJPVgoJYk\nSZJ6MFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerBQC1JkiT1YKCWJEmS\nejBQS5IkST0YqCVJkqQeDNSSJElSD4MG6iQHJ7k+yZokJ06w/6gk65Nc2f78zpD1SJIkSTNtm6E6\nTrIAOB14PrAOWJlkeVVdO67pOVV13FB1SJIkSUMacoR6X2BNVd1QVXcDZwOHDXh9kiRJ0mY3ZKDe\nDVjbWV/Xbhvvt5JcleT8JLsPWI8kSZI044YM1JlgW41bvwBYUlVPBj4LnDlhR8kxSVYlWbV+/foZ\nLlOSJEmaviED9TqgO+K8GLi526Cqbq2qu9rVDwJPn6ijqjqjqpZV1bJFixYNUqwkSZI0HUMG6pXA\nPkn2SrIdcASwvNsgySM6q4cC1w1YjyRJkjTjBjvLR1VtSHIccCGwAPhwVa1OciqwqqqWA8cnORTY\nAHwfOGqoeiRJkqQhDBaoAapqBbBi3LaTO8tvAt40ZA2SJEnSkPymREmSJKkHA7UkSZLUg4FakiRJ\n6sFALUmSJPVgoJYkSZJ6MFBLkiRJPRioJUmSpB6mDNRJjkvy0M1RjCRJkjTfjDJC/YvAyiTnJjk4\nSYYuSpIkSZovpgzUVXUSsA/wIZqvBv96krcm2Xvg2iRJkqQ5b6Q51FVVwHfbnw3AQ4Hzk5w2YG2S\nJEnSnLfNVA2SHA+8ErgF+FvgD6rqniQPAr4OvHHYEiVJkqS5a8pADSwEXlRVN3U3VtV9SX5jmLIk\nSZKk+WGUKR8rgO+PrSTZKcl+AFV13VCFSZIkSfPBKIH6b4A7O+s/bLdJkiRJW71RAnXaDyUCzVQP\nRpsqIkmSJG3xRgnUNyQ5Psm27c9rgRuGLkySJEmaD0YJ1McC+wPfBtYB+wHHDFmUJEmSNF9MOXWj\nqr4HHLEZapEkSZLmnVHOQ70DcDTwBGCHse1V9T8HrEuSJEmaF0aZ8vFR4BeBXwP+FVgM3DFkUZIk\nSdJ8MUqgfnRV/THww6o6E/h14EnDliVJkiTND6ME6nva37cneSKwC7BksIokSZKkeWSU80mfkeSh\nwEnAcmBH4I8HrUqSJEmaJzYaqJM8CPivqroN+DfgUZulKkmSJGme2OiUj/ZbEY/bTLVIkiRJ884o\nc6j/JckJSXZP8rCxn8ErkyRJkuaBUeZQj51v+tWdbYXTPyRJkqSRvilxr81RiCRJkjQfjfJNia+Y\naHtV/f3MlyNJkiTNL6NM+XhGZ3kH4EDgS4CBWpIkSVu9UaZ8vKa7nmQXmq8jn1KSg4G/BBYAf1tV\nb5uk3YuB84BnVNWqUfqWJEmS5oJRzvIx3o+AfaZqlGQBcDpwCLAUODLJ0gna7QQcD3xxGrVIkiRJ\ns2qUOdQX0JzVA5oAvhQ4d4S+9wXWVNUNbT9nA4cB145r96fAacAJI9YsSZIkzRmjzKH+i87yBuCm\nqlo3wuV2A9Z21tcB+3UbJHkasHtVfSrJpIE6yTHAMQB77LHHCFctSZIkbR6jBOpvAd+pqp8AJHlw\nkiVVdeMUl8sE2+qnO5uvNX83cNRUBVTVGcAZAMuWLaspmkuSJEmbzShzqM8D7uus39tum8o6YPfO\n+mLg5s76TsATgc8luRF4JrA8ybIR+pYkSZLmhFEC9TZVdffYSru83QiXWwnsk2SvJNsBRwDLO/38\noKoWVtWSqloCXA4c6lk+JEmSNJ+MEqjXJzl0bCXJYcAtU12oqjYAxwEXAtcB51bV6iSndvuTJEmS\n5rNR5lAfC3wsyXvb9XXAhN+eOF5VrQBWjNt28iRtDxilT0mSJGkuGeWLXb4BPDPJjkCq6o7hy5Ik\nSZLmhymnfCR5a5Jdq+rOqrojyUOT/NnmKE6SJEma60aZQ31IVd0+tlJVtwEvGK4kSZIkaf4YJVAv\nSLL92EqSBwPbb6S9JEmStNUY5UOJ/wBcnOQj7fqrgDOHK0mSJEmaP0b5UOJpSa4CDqL59sPPAHsO\nXZgkSZI0H4wy5QPguzTflvhbwIE055WWJEmStnqTjlAneQzNtxseCdwKnENz2rznbabaJEmSpDlv\nY1M+vgr8O/Dfq2oNQJLXb5aqJEmSpHliY1M+fotmqselST6Y5ECaOdSSJEmSWpMG6qr6RFUdDjwO\n+BzweuDhSf4mya9upvokSZKkOW3KDyVW1Q+r6mNV9RvAYuBK4MTBK5MkSZLmgVHP8gFAVX2/qj5Q\nVb8yVEGSJEnSfLJJgVqSJEnS/RmoJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIkSerBQC1J\nkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSDgVqSJEnqwUAtSZIk\n9WCgliRJknowUEuSJEk9DBqokxyc5Poka5KcOMH+Y5NcneTKJJ9PsnTIeiRJkqSZNligTrIAOB04\nBFgKHDlBYD6rqp5UVU8FTgPeNVQ9kiRJ0hCGHKHeF1hTVTdU1d3A2cBh3QZV9V+d1Z8DasB6JEmS\npBm3zYB97was7ayvA/Yb3yjJq4E3ANsBvzJRR0mOAY4B2GOPPWa8UEmSJGm6hhyhzgTbHjACXVWn\nV9XewB8CJ03UUVWdUVXLqmrZokWLZrhMSZIkafqGDNTrgN0764uBmzfS/mzghQPWI0mSJM24IQP1\nSmCfJHsl2Q44AljebZBkn87qrwNfH7AeSZIkacYNNoe6qjYkOQ64EFgAfLiqVic5FVhVVcuB45Ic\nBNwD3Aa8cqh6JEmSpCEM+aFEqmoFsGLctpM7y68d8volSZKkoflNiZIkSVIPBmpJkiSpBwO1JEmS\n1IOBWpIkSerBQC1JkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSD\ngVqSJEnqwUAtSZIk9WCgliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4Fa\nkiRJ6sFALUmSJPVgoJYkSZJ6MFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIk\nSerBQC1JkiT1YKCWJEmSehg0UCc5OMn1SdYkOXGC/W9Icm2Sq5JcnGTPIeuRJEmSZtpggTrJAuB0\n4BBgKXBkkqXjmn0ZWFZVTwbOB04bqh5JkiRpCEOOUO8LrKmqG6rqbuBs4LBug6q6tKp+1K5eDiwe\nsB5JkiRpxg0ZqHcD1nbW17XbJnM08OmJdiQ5JsmqJKvWr18/gyVKkiRJ/QwZqDPBtpqwYfJyYBnw\njon2V9UZVbWsqpYtWrRoBkuUJEmS+tlmwL7XAbt31hcDN49vlOQg4M3Ac6vqrgHrkSRJkmbckCPU\nK4F9kuyVZDvgCGB5t0GSpwEfAA6tqu8NWIskSZI0iMECdVVtAI4DLgSuA86tqtVJTk1yaNvsHcCO\nwHlJrkyyfJLuJEmSpDlpyCkfVNUKYMW4bSd3lg8a8volSZKkoflNiZIkSVIPBmpJkiSpBwO1JEmS\n1IOBWpIkSerBQC1JkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSD\ngVqSJEnqwUAtSZIk9WCgliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4Fa\nkiRJ6sFALUmSJPVgoJYkSZJ6MFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS1IOBWpIk\nSerBQC1JkiT1YKCWJEmSehg0UCc5OMn1SdYkOXGC/b+c5EtJNiR58ZC1SJIkSUMYLFAnWQCcDhwC\nLAWOTLJ0XLNvAUcBZw1VhyRJkjSkbQbse19gTVXdAJDkbOAw4NqxBlV1Y7vvvgHrkCRJkgYz5JSP\n3YC1nfV17TZJkiRpizFkoM4E22paHSXHJFmVZNX69et7liVJkiTNnCED9Tpg9876YuDm6XRUVWdU\n1bKqWrZo0aIZKU6SJEmaCUMG6pXAPkn2SrIdcASwfMDrkyRJkja7wQJ1VW0AjgMuBK4Dzq2q1UlO\nTXIoQJJnJFkHvAT4QJLVQ9UjSZIkDWHIs3xQVSuAFeO2ndxZXkkzFUSSJEmal/ymREmSJKkHA7Uk\nSZLUg4FakiRJ6sFALUmSJPVgoJYkSZJ6MFBLkiRJPRioJUmSpB4M1JIkSVIPBmpJkiSpBwO1JEmS\n1IOBWpIkSerBQC1JkiT1YKCWJEmSejBQS5IkST0YqCVJkqQeDNSSJElSDwZqSZIkqQcDtSRJktSD\ngVqSJEnqwUAtSZIk9WCgliRJknowUEuSJEk9GKglSZKkHgzUkiRJUg8GakmSJKkHA7UkSZLUg4Fa\nkiRJ6sFALUmSJPVgoJYkSZJ6MFBLkiRJPQwaqJMcnOT6JGuSnDjB/u2TnNPu/2KSJUPWI0mSJM20\nwQJ1kgXA6cAhwFLgyCRLxzU7Gritqh4NvBt4+1D1SJIkSUMYcoR6X2BNVd1QVXcDZwOHjWtzGHBm\nu3w+cGCSDFiTJEmSNKOGDNS7AWs76+vabRO2qaoNwA+Anx+wJkmSJGlGbTNg3xONNNc02pDkGOCY\ndvXOJNf3rG1LthC4ZbaLmNLLtuo3IjxGc5/HaO7zGM19HqO5z2M0tT1HaTRkoF4H7N5ZXwzcPEmb\ndUm2AXYBvj++o6o6AzhjoDq3KElWVdWy2a5Dk/MYzX0eo7nPYzT3eYzmPo/RzBlyysdKYJ8keyXZ\nDjgCWD6uzXLgle3yi4FLquoBI9SSJEnSXDXYCHVVbUhyHHAhsAD4cFWtTnIqsKqqlgMfAj6aZA3N\nyPQRQ9UjSZIkDWHIKR9U1QpgxbhtJ3eWfwK8ZMgatkJOjZn7PEZzn8do7vMYzX0eo7nPYzRD4gwL\nSZIkafr86nFJkiSpBwP1NCX5xSRnJ/lGkmuTrEjymBns/4Ak+3eWLxu3f5sk/5nkEZvY765J/vc0\na/p/07ncbEty52zX0DX+GCR5ZJLzZ6ufrVmSSvLOzvoJSU5pl09J8qMkv9DZP6ceS3NZkjcnWZ3k\nqiRXJtkvyeuSPGQz13FskldszuucD2b6+CR5SpIrO+tHtn8/27brT0pyVY96lyS5ZrqXl4ZmoJ6G\n9tscPwF8rqr2rqqlwB8BD5/BqzkA2L9d/jdgcZIlnf0HAddU1Xc2sd9dgU0K1O3XyFNV+0/VdtS+\ntnL3OwZVdXNVvXgW+9ma3QW8KMnCSfbfAvz+Zqxni5DkWcBvAP+tqp5M83y1FngdsFkDdVW9v6r+\nfnNe51w30PG5GtgzyU7t+v7AV4Gndda/sAk1+loxkPY0xZphBurpeR5wT1W9f2xDVV0JfD7JO5Jc\nk+TqJIfDT0eYPzXWNsl7kxzVLt+Y5C1JvtRe5nFtcD4WeH37H/+zgfOAwzs1HAF8vO1j7ySfSXJF\nkn9P8rh2+8OTfCLJV9qf/YG3AXu3IxLvSGOymi9NchbNE+VPR+eSnNpe/sok307ykXb7y5P8R7v9\nA2NPiEnubC/zReBZM3kgpivJoiT/mGRl+/PsdvspSc5MclF7bF6U5LT2vvlMZ7TlxiRvTXJZklVJ\n/luSC9O8Y3Fs22bHJBd3ju1h7dWPPwY/HXlJ8red+3Z9kj+ZZj87JPlI2/7LSZ7Xbj8qyT+1t+Xr\nSU7bfPf6nLSB5kM5r59k/4eBw5M8bPOVtEV4BHBLVd0FUFW30Jwa9ZHApUkuBUjyq+3f0JeSnJdk\nx3b7KH9fByT51yTnJvlakrcleVn7HHR1kr3bdqckOaFd/lySt7dtvpbkl9rtD2n7uSrJOUm+mGRL\nPjfvqMfnyPa+vCbJ28cu3D6nv7M9bhcnWVRV99GcLne/ttnTgdP52cDQ/sD/ay9/YPu8dHWSDyfZ\nvt1+Y5KTk3weeEmSp6d57boMeHXn+p/Qea25Ksk+g91Tc1T7fH9dkg+meafhoiQPTvKM9j65rH1d\nGHtNOKr9G7sAuGiy15W236+2r0XXJPlYkoOSfKF9zdi3bffczmvVl/Ozf6S2XlXlzyb+AMcD755g\n+28B/0JzmsCHA9+ieeI6APhUp917gaPa5RuB17TL/xv423b5FOCEzmWeAXy5Xd4e+B7w0Hb9YmCf\ndnk/mvN5A5wDvK5dXkDzxTn4J+8DAAAHoUlEQVRLaEa2R6n5h8BenbZ3jru9uwBX0TxxPh64ANi2\n3fc+4BXtcgG/PYvH684Jtp0FPKdd3gO4rnO/fx7YFngK8CPgkHbfJ4AXdo7b77XL727vh52ARcD3\n2u3bADu3ywuBNTTfDjr+GNxvvd22J83ozp7T6YdmVPUj7fLj2uO6A3AUcEN77HYAbgJ2n+2/qdl8\nbAA7t8dzF+AE4JTOY+EE4GTgLZM9lvyZ8H7dEbgS+Fr7XPDcdvuNwMJ2eSHNu28/167/IXByp91U\nf18HALfTPF9tD3y7c5xeC7ynexzb5c8B72yXXwB8tl0+AfhAu/xEmn+0ls32/TjLx+eR7fPGovY5\n6BJ+9vxXwMva5ZOB93bu65OBn6N5Ht0bOLfd93XgUe3zzlrgMe32v+dnr1M3Am/s1HlVp7Z38LPn\nt7/uXP92wINn+z6dhWO4pH2cPrVdPxd4OXANsH+77W2d++womi/Te1i7vrHXlQ3Ak2gGXa+gGVgI\ncBjwyfYyFwDP7jyetpnt+2S2fxyhnlnPAT5eVfdW1X8C/0oThKfyT+3vK2gezA9QVSuBHZM8FjgE\nuLyqbmtHdPYHzkszmv0BmhcYgF8B/qa9/L1V9YNNrPk/quqbE9WTJMDHaP6xuAI4kCZYr2zrOJDm\nyRPgXuAfp74bNquDgPe2tS4Hdu78h/3pqrqHZmR+AfCZdvvV3P/4LO9s/2JV3VFV64GfJNmV5gno\nrWnmDX4W2I0RpgUl2YHmHYnjquqmafbzHOCjAFX1VZrgPDbH/+Kq+kE1p628lhG/VnVLVVX/RfOi\nfvwkTf4KeGWSnTdfVfNbVd1J83xwDLAeOCftu3IdzwSWAl9o/w5fyf0fi1P9fQGsrKrvVDPS+g3g\nos5llkxS3kTPt88Bzm5rv4YmyG2xRjw+z6CZ1ri+qjbQPN//crvvPpoBG4B/oLn/oJnSsT+wL82x\n+Qbw6CSLgB2r6gbgscA3q+pr7WXO7PTLWL9JdgF2rap/bbd/tNPmMuCPkvwhsGdV/Xgad8OW4JvV\nvDsOP3s871RVY593Omtc+3+pqrFvo97Y68o3q+rqat51WE3zmlHc/+/qC8C7khxPc5w2zOxNm3+c\nRzM9q2neHhtvsi+b38D9p9fsMG7/Xe3ve9n4MTmbZqrH42mne7T93l5VT91YwRsxWc3QjFBP5hRg\nXVV9pNPPmVX1pgna/qSq7p1mfUN5EPCs8U/Ezf8JjL0Nel+Se9onEmheRLrH567O9rs628favYxm\ndOfpVXVPkht54LGfyPuBf6qqz7br0+lnY8e1W+tUj7mtxXuALwEfGb+jqm5PM/VpWh/m3Vq1f/Of\nAz6X5Gp+9q24Y0LzAn/kJF1M9ffFBNvvmqDNZP12H/sb+3vZIo14fEburv19OU0Qfw5N6IVmVPQI\n2ukeI/Q79rqTTr/3v7Kqs9JMIfx14MIkv1NVl2xCvVuK8c/lj5yiffc1fWOvK1P+XVXV25L8X5p3\nei5PclA7eLPVcoR6ei4Btk/yu2MbkjwDuI1mvuWC9j/yXwb+g2Z0cGmS7dv/ug8c4TruoHmLs+vj\nNG/p/Art6E07uvbNJC9p60iSp7TtLwZ+r92+oB1hG9/vv01S86SS/AbwfO4/oncx8OK0Z0RI8rAk\nc3nk8yLguLGVJNP9h2RjdqF5e/qeNHOYx+6PiY7tWB2vphlheFuffmiO68vaPh9DM63l+j43ZkvW\njtqcCxw9SZN3Af8L//kYSZLHjpvX+lSa58HuY/Zy4NlJHt1e5iGZwTMlbaLPA7/d1rGU5u3uLdaI\nx+eLwHOTLEzzeZgjad7BhCY7jA0qvZTm/qOq7qCZznEUPwvUl9F82HEsUH8VWDJ23IH/0en3p6rq\nduAHScZGv1/Wqf9RwA1V9Vc0r4VP3pTbvwW7DbgjyTPb9Y19+/RkrysjSbJ3O4r9dmAVzdTCrZqB\nehraEcvfBJ6f5kMyq2lGbM+ieavwKzSh+41V9d2qWkvzYn0VzdtmXx7hai4AfrOd8P9L7fVeSzOn\n95KqGv+f5tFJvkIzej72obXXAs9rRx+uAJ5QVbfSvMV6TZJ30MwLfkDNU9T2+zT/CY99KOTUtraT\naD7scBXNvOxNOqXfgB6SZF3n5w00/wwsaz+8cS3Nh0Bn2sfa61hFc4y+CjDBMeg6AXhS58Mex06z\nn/cBC9pjfw7NnP270Ma8k2Yu4QNU86GtT9DM1dXUdgTOTHNK0atopnacQvMB0E8nubSdvnEU8PG2\nzeXM3ovy+4BFbR1/SPOcONEUuS3FKMfnO8CbgEtpXh++VFX/3F7+h8ATklxBM8BzaqfvLwDbt697\n0ATqR9EG6naq2atopileTTPq+X4m9irg9DQfSuy+m3g4cE07VehxNFO21DgaOKO9z8Lkj+MJX1c2\nweva156v0BybT0+34C2F35QoSdqqtSOw21bVT9KcHeRimg/N3T3Lpc1JSe6sqh1nuw49UJId2zny\nJDkReERVvXaWy9oq+PalJGlr9xCa08VtSzOq93uGac1Tv57kTTT57iaad4G0GThCLUmSJPXgHGpJ\nkiSpBwO1JEmS1IOBWpIkSerBQC1JkiT1YKCWJEmSejBQS5IkST38f46h1BSdNfQOAAAAAElFTkSu\nQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x1162fcda0>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "6HH0ld3Breqe",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "The best model is NN with  better performance than the others with accuracy 0.79."
      ]
    },
    {
      "metadata": {
        "id": "ocljxoDAroQx",
        "colab_type": "code",
        "colab": {
          "autoexec": {
            "startup": false,
            "wait_interval": 0
          }
        }
      },
      "cell_type": "code",
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}