Progetto_Data_Intensive.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Progetto Data Intensive.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "hInOIDeEzOlU",
        "RaxUJRhpzTFf",
        "9A3VyoXEB5DX"
      ],
      "authorship_tag": "ABX9TyNWv646gfQF4yMWLEUpxnbQ",
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/massone99/Data-science-project/blob/main/Progetto_Data_Intensive.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C9IE1eu-hMOK"
      },
      "source": [
        "# Determinare la predisposizione di un individuo a colpi apoplettici (Ictus)\n",
        "\n",
        "**Programmazione di Applicazioni Data Intensive**  \n",
        "Laurea in Ingegneria e Scienze Informatiche  \n",
        "DISI - Università di Bologna, Cesena\n",
        "\n",
        "Lorenzo Massone"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Mb-noTsBhRG5"
      },
      "source": [
        "##Descrizione del problema e analisi esplorativa\n",
        "\n",
        "Si deve realizzare un modello che sia in grado di classificare individui come a rischio apoplettico o meno, in base ai dati forniti riguardo il loro attuale stato di salute.\n",
        "\n",
        "Vengono importate le librerie necessarie per scaricare i file, organizzare le strutture dati e disegnare i grafici."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SM8aea_E24NY",
        "outputId": "57ca81bd-969d-4284-d5c3-e46d7c094795"
      },
      "source": [
        "pip install seaborn"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: seaborn in /usr/local/lib/python3.7/dist-packages (0.11.1)\n",
            "Requirement already satisfied: pandas>=0.23 in /usr/local/lib/python3.7/dist-packages (from seaborn) (1.1.5)\n",
            "Requirement already satisfied: matplotlib>=2.2 in /usr/local/lib/python3.7/dist-packages (from seaborn) (3.2.2)\n",
            "Requirement already satisfied: scipy>=1.0 in /usr/local/lib/python3.7/dist-packages (from seaborn) (1.4.1)\n",
            "Requirement already satisfied: numpy>=1.15 in /usr/local/lib/python3.7/dist-packages (from seaborn) (1.19.5)\n",
            "Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.23->seaborn) (2018.9)\n",
            "Requirement already satisfied: python-dateutil>=2.7.3 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.23->seaborn) (2.8.1)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib>=2.2->seaborn) (2.4.7)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib>=2.2->seaborn) (1.3.1)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib>=2.2->seaborn) (0.10.0)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.7.3->pandas>=0.23->seaborn) (1.15.0)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-SmSh0s1hDup"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns \n",
        "import sklearn\n",
        "%matplotlib inline"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bNUDKDCcj9wK"
      },
      "source": [
        "### Caricamento dei dati e preprocessing"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "50NsoV99UZpT"
      },
      "source": [
        "Scarichiamo i dati che andremo ad utilizzare per addestrare i nostri modelli:\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E_NP32nEhRG6"
      },
      "source": [
        "import os\n",
        "\n",
        "file_zip_name = \"stroke.zip\"\n",
        "file_csv_name = \"healthcare-dataset-stroke-data.csv\"\n",
        "\n",
        "if not os.path.exists(file_zip_name):\n",
        "    from urllib.request import urlretrieve\n",
        "    urlretrieve(\"https://git.io/JnEMm\", file_zip_name)\n",
        "    from zipfile import ZipFile"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RXKDQ-CmE4n7",
        "outputId": "4362fcb9-e1ce-4b14-8f68-c80db69c0747"
      },
      "source": [
        "!unzip \"stroke.zip\""
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Archive:  stroke.zip\n",
            "  inflating: healthcare-dataset-stroke-data.csv  \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xlJN4NmCVXqc"
      },
      "source": [
        "data_raw = pd.read_csv(file_csv_name)"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "MdH_mHdofQEX",
        "outputId": "f4999913-e09e-4728-8fd3-4b7c5dbfa6b5"
      },
      "source": [
        "data_raw.head()"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>gender</th>\n",
              "      <th>age</th>\n",
              "      <th>hypertension</th>\n",
              "      <th>heart_disease</th>\n",
              "      <th>ever_married</th>\n",
              "      <th>work_type</th>\n",
              "      <th>Residence_type</th>\n",
              "      <th>avg_glucose_level</th>\n",
              "      <th>bmi</th>\n",
              "      <th>smoking_status</th>\n",
              "      <th>stroke</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>9046</td>\n",
              "      <td>Male</td>\n",
              "      <td>67.0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>Yes</td>\n",
              "      <td>Private</td>\n",
              "      <td>Urban</td>\n",
              "      <td>228.69</td>\n",
              "      <td>36.6</td>\n",
              "      <td>formerly smoked</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>51676</td>\n",
              "      <td>Female</td>\n",
              "      <td>61.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>Yes</td>\n",
              "      <td>Self-employed</td>\n",
              "      <td>Rural</td>\n",
              "      <td>202.21</td>\n",
              "      <td>NaN</td>\n",
              "      <td>never smoked</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>31112</td>\n",
              "      <td>Male</td>\n",
              "      <td>80.0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>Yes</td>\n",
              "      <td>Private</td>\n",
              "      <td>Rural</td>\n",
              "      <td>105.92</td>\n",
              "      <td>32.5</td>\n",
              "      <td>never smoked</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>60182</td>\n",
              "      <td>Female</td>\n",
              "      <td>49.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>Yes</td>\n",
              "      <td>Private</td>\n",
              "      <td>Urban</td>\n",
              "      <td>171.23</td>\n",
              "      <td>34.4</td>\n",
              "      <td>smokes</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1665</td>\n",
              "      <td>Female</td>\n",
              "      <td>79.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>Yes</td>\n",
              "      <td>Self-employed</td>\n",
              "      <td>Rural</td>\n",
              "      <td>174.12</td>\n",
              "      <td>24.0</td>\n",
              "      <td>never smoked</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      id  gender   age  ...   bmi   smoking_status stroke\n",
              "0   9046    Male  67.0  ...  36.6  formerly smoked      1\n",
              "1  51676  Female  61.0  ...   NaN     never smoked      1\n",
              "2  31112    Male  80.0  ...  32.5     never smoked      1\n",
              "3  60182  Female  49.0  ...  34.4           smokes      1\n",
              "4   1665  Female  79.0  ...  24.0     never smoked      1\n",
              "\n",
              "[5 rows x 12 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6iJJcC7Kudkq"
      },
      "source": [
        "Di seguito sono riportate le principali informazioni riguardo i dati, tra cui le dimensioni in memoria, il numero di istanze non nulle e i tipi delle feature presenti nelle varie istanze del dataset."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1iTUsHdYVhFr",
        "outputId": "3fc6cc4e-34e3-4e28-8dcf-85cd3bba8736"
      },
      "source": [
        "data_raw.info(memory_usage=\"deep\")"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 5110 entries, 0 to 5109\n",
            "Data columns (total 12 columns):\n",
            " #   Column             Non-Null Count  Dtype  \n",
            "---  ------             --------------  -----  \n",
            " 0   id                 5110 non-null   int64  \n",
            " 1   gender             5110 non-null   object \n",
            " 2   age                5110 non-null   float64\n",
            " 3   hypertension       5110 non-null   int64  \n",
            " 4   heart_disease      5110 non-null   int64  \n",
            " 5   ever_married       5110 non-null   object \n",
            " 6   work_type          5110 non-null   object \n",
            " 7   Residence_type     5110 non-null   object \n",
            " 8   avg_glucose_level  5110 non-null   float64\n",
            " 9   bmi                4909 non-null   float64\n",
            " 10  smoking_status     5110 non-null   object \n",
            " 11  stroke             5110 non-null   int64  \n",
            "dtypes: float64(3), int64(4), object(5)\n",
            "memory usage: 1.8 MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f71dvasfay0d"
      },
      "source": [
        "###Descrizione delle feature\n",
        "Il DataFrame creato presenta le seguenti feature:\n",
        "- **`gender`** sesso dell'individuo\n",
        "  - **`Male`**\n",
        "  - **`Female`**\n",
        "\n",
        "- **`age`** età dell'individuo\n",
        "\n",
        "- **`hypertension`** indica se il paziente soffre di ipertensione\n",
        "  - 0: caso negativo\n",
        "  - 1: caso affermativo\n",
        "\n",
        "- **`heart_disease`** indica la presenza di malattie cardiache\n",
        "  - 0: nessuna malattia\n",
        "  - 1: ha almeno una malattia cardiaca\n",
        "\n",
        "- **`ever_married`** se l'individuo si è mai sposato\n",
        "\n",
        "- **`work_type`** indica il tipo di lavoro dell'individuo\n",
        "  - **`children`**\n",
        "  - **`Govt_jov`**\n",
        "  - **`Never_worked`**\n",
        "  - **`Private`**\n",
        "  - **`Self-employed`**\n",
        "\n",
        "- **`Residence_type`** indica il tipo di abitazione dell'individuo\n",
        "  - **`Rural`**\n",
        "  - **`Urban`**\n",
        " \n",
        "- **`avg_glucose_level`** livello medio di glucosio nel sangue\n",
        "\n",
        "- **`bmi`** indice di massa corporea\n",
        "\n",
        "- **`smoking_status`** indica se l'individuo è un fumatore\n",
        "  - **`formerly smoked`**: precedentemente fumatore\n",
        "  - **`never smoked`**: non ha mai fumato\n",
        "  - **`smokes`**: fumatore\n",
        "  - **`Unknown`**: non noto\n",
        "\n",
        "- **`stroke`** indica se l'individuo ha già avuto un colpo apoplettico\n",
        "  - 0 = caso negativo\n",
        "  - 1 = caso positivo\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0QrMAb1BpuLL"
      },
      "source": [
        "Possiamo notare come tutte le feature categoriche presenti nel DataFrame siano memorizzate con tipo `Object`. Procediamo quindi a convertire le seguenti feature nel tipo categorico:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gEdRRvdwu1K3",
        "outputId": "bb9418e4-4fe2-4daa-b599-f1bd1342fa8e"
      },
      "source": [
        "categorical = [\"smoking_status\", \"Residence_type\", \"gender\",];\n",
        "data = pd.read_csv(file_csv_name, dtype={x:\"category\" for x in categorical});\n",
        "data.info(memory_usage=\"deep\")"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 5110 entries, 0 to 5109\n",
            "Data columns (total 12 columns):\n",
            " #   Column             Non-Null Count  Dtype   \n",
            "---  ------             --------------  -----   \n",
            " 0   id                 5110 non-null   int64   \n",
            " 1   gender             5110 non-null   category\n",
            " 2   age                5110 non-null   float64 \n",
            " 3   hypertension       5110 non-null   int64   \n",
            " 4   heart_disease      5110 non-null   int64   \n",
            " 5   ever_married       5110 non-null   object  \n",
            " 6   work_type          5110 non-null   object  \n",
            " 7   Residence_type     5110 non-null   category\n",
            " 8   avg_glucose_level  5110 non-null   float64 \n",
            " 9   bmi                4909 non-null   float64 \n",
            " 10  smoking_status     5110 non-null   category\n",
            " 11  stroke             5110 non-null   int64   \n",
            "dtypes: category(3), float64(3), int64(4), object(2)\n",
            "memory usage: 918.7 KB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yKecTQXq2FOs"
      },
      "source": [
        "Si può notare come tale conversione comporti un uso di memoria quasi dimezzato.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gtcuvJM-2TIT"
      },
      "source": [
        "Procediamo ora alla rimozione della colonna degli `id` non rilevanti nello sviluppo del nostro modello, in modo tale da risparmiare ulteriormente lo spazio a nostra disposizione."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_rKm_miU1oTf"
      },
      "source": [
        "data.drop(columns=[\"id\"], inplace=True)"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n-xCVwHXHeG6"
      },
      "source": [
        "###Analisi generale dei dati\n",
        "Studiamo adesso le caratteristiche dei dati in nostro possesso per meglio comprendere le caratteristiche del dominio applicativo in questione"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        },
        "id": "dRrbzM85IV2F",
        "outputId": "6e1744e3-93ac-48bd-ee34-1c080a1b3f0b"
      },
      "source": [
        "data.describe()"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>age</th>\n",
              "      <th>hypertension</th>\n",
              "      <th>heart_disease</th>\n",
              "      <th>avg_glucose_level</th>\n",
              "      <th>bmi</th>\n",
              "      <th>stroke</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>5110.000000</td>\n",
              "      <td>5110.000000</td>\n",
              "      <td>5110.000000</td>\n",
              "      <td>5110.000000</td>\n",
              "      <td>4909.000000</td>\n",
              "      <td>5110.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>43.226614</td>\n",
              "      <td>0.097456</td>\n",
              "      <td>0.054012</td>\n",
              "      <td>106.147677</td>\n",
              "      <td>28.893237</td>\n",
              "      <td>0.048728</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>22.612647</td>\n",
              "      <td>0.296607</td>\n",
              "      <td>0.226063</td>\n",
              "      <td>45.283560</td>\n",
              "      <td>7.854067</td>\n",
              "      <td>0.215320</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>0.080000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>55.120000</td>\n",
              "      <td>10.300000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>25.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>77.245000</td>\n",
              "      <td>23.500000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>45.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>91.885000</td>\n",
              "      <td>28.100000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>61.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>114.090000</td>\n",
              "      <td>33.100000</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>82.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>271.740000</td>\n",
              "      <td>97.600000</td>\n",
              "      <td>1.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "               age  hypertension  ...          bmi       stroke\n",
              "count  5110.000000   5110.000000  ...  4909.000000  5110.000000\n",
              "mean     43.226614      0.097456  ...    28.893237     0.048728\n",
              "std      22.612647      0.296607  ...     7.854067     0.215320\n",
              "min       0.080000      0.000000  ...    10.300000     0.000000\n",
              "25%      25.000000      0.000000  ...    23.500000     0.000000\n",
              "50%      45.000000      0.000000  ...    28.100000     0.000000\n",
              "75%      61.000000      0.000000  ...    33.100000     0.000000\n",
              "max      82.000000      1.000000  ...    97.600000     1.000000\n",
              "\n",
              "[8 rows x 6 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x57d-FnLIgRW"
      },
      "source": [
        "Si può notare come solamente un terzo delle righe sia già stato soggetto a colpi apoplettici (per essere precisi gli individui che rientrano nella fascia di età più alta) comportando un forte sbilanciamento delle classi, che andrà necessariamente compensato più avanti per non ridurre l'accuratezza del modello.\n",
        "Tutto ciò è reso ancora più palese rappresentando la distribuzione di diagnosi negative e diagnosi positive in un grafico a torta."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 248
        },
        "id": "G06Fzt4feKdp",
        "outputId": "ec7c5587-83ad-44ba-f24e-fb1e54d7f475"
      },
      "source": [
        "data.stroke.value_counts().plot.pie(autopct=\"%.1f%%\");"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BtoXh-_hRYvJ"
      },
      "source": [
        "Studiamo ora la distribuzione delle feature principali rappresentandole per via grafica."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iopnFPP21J5a"
      },
      "source": [
        "Rappresentiamo inizialmente in che modo malattie cardiache e ipertensione siano collegate alla possibilità di colpi apoplettici:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "ZU40FpZmRf_q",
        "outputId": "297e4bcb-349e-402e-aa22-c8623298a8a3"
      },
      "source": [
        "data.age.plot.hist(bins=40);"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Filn1xpUCTLV",
        "outputId": "9e30abed-4308-466a-c3bd-b78e96e2bd4a"
      },
      "source": [
        "data[\"age\"].describe()"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "count    5110.000000\n",
              "mean       43.226614\n",
              "std        22.612647\n",
              "min         0.080000\n",
              "25%        25.000000\n",
              "50%        45.000000\n",
              "75%        61.000000\n",
              "max        82.000000\n",
              "Name: age, dtype: float64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KySRoB34aUKc"
      },
      "source": [
        "La distribuzione delle età degli individui sotto osservazione risulta abbastanza omogenea come si può notare dal grafico e dalla deviazione standard non troppo elevata."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1LlgRijkC7NX"
      },
      "source": [
        "Confrontiamo ora il sesso degli individui sotto esame:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 264
        },
        "id": "xE3_vIdiDBfd",
        "outputId": "af7137cb-4ab1-409d-e8d8-d00985ae2e94"
      },
      "source": [
        "data.gender.value_counts().plot.pie(autopct=\"%.1f%%\", colors= [\"pink\", \"skyblue\"]);\n",
        "plt.title(\"Distribuzione sesso degli individui\");\n",
        "plt.ylabel(\"\");"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "J4He0XanE61D",
        "outputId": "aaacd719-333e-4787-ed72-db3bbc90eab2"
      },
      "source": [
        "data.gender.value_counts()"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Female    2994\n",
              "Male      2115\n",
              "Other        1\n",
              "Name: gender, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OAlhRCSADWt-"
      },
      "source": [
        "Come si può notare la maggior parte degli individui sotto esame sono di sesso femminile. Notiamo anche la presenza di una terza categoria `Other` che in questo caso risulta poco significativa in quanto rappresentata da un'unica istanza."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ua-kLjtsFCJz"
      },
      "source": [
        "Approfondiamo  visualizzando l'incidenza di ictus sui vari generi"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "id": "RS6hChl6D3_G",
        "outputId": "e46e5eb0-b9bc-42f1-e06a-971ffd06043b"
      },
      "source": [
        "sns.set_theme(style=\"whitegrid\")\n",
        "plt.title(\"Incidenza di ictus a confronto con il sesso dell'individuo\")\n",
        "sns.barplot(x=\"gender\", y=\"stroke\", data=data);"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xrsoI0-yxQ2w"
      },
      "source": [
        "Negli uomini l'incidenza di colpi apoplettici risulta maggiore rispetto alle donne. Si nota anche la presenza di un terzo valore `Other` con cardinalità molto bassa. Procediamo a rimuoverlo:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E4GqDkqExsnq"
      },
      "source": [
        "data = data[data.gender != \"Other\"]"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "id": "6HvOwWRVZ1zz",
        "outputId": "01d9ed8c-3b6c-4f50-9d8c-2c6756655675"
      },
      "source": [
        "data.avg_glucose_level.plot.hist(bins=40, color=\"coral\");\n",
        "plt.title(\"Distribuzione glicemia negli individui sotto osservazione\")\n",
        "plt.xlabel(\"Avg Glucose\");"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BCAdPTqveBmL",
        "outputId": "d71ca1d3-a0a9-463a-b03e-ed2582e85f95"
      },
      "source": [
        "data.avg_glucose_level.std()/data.avg_glucose_level.mean()"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.4266519060260945"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3z6DRdgTapJP"
      },
      "source": [
        "La maggior parte degli individui in questione presenta un tasso di glicemia ordinario (ovvero compreso nell'intervallo tra i 70 e i 100 mg/dL) anche se la deviazione standard normalizzata non è indifferente in quanto ammonta circa a 0.42. \n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MGwRVfWHI9el"
      },
      "source": [
        "Procediamo rappresentando come si correli il matrimonio all'incidenza di ictus:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "8dMCbvwToKL5",
        "outputId": "2f23f504-d7e6-48c4-ea19-dd13f4d52bca"
      },
      "source": [
        "sns.barplot(x=\"ever_married\", y=\"stroke\", data=data);"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1HXM6sQsJjCn"
      },
      "source": [
        "Il grafico sembra suggerire che gli individui che si sono sposati siano maggiormente propensi a colpi apoplettici, tuttavia ciò non risulta vero in quanto gli individui sposati sono statisticamente più rappresentati rispetto agli scapoli:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 264
        },
        "id": "HEgAieRqKV3k",
        "outputId": "c103d278-263f-42e7-9432-8f0a0ff95c99"
      },
      "source": [
        "data.ever_married.value_counts().plot.pie(autopct=\"%.1f%%\", colors= [\"orange\", \"yellow\"]);\n",
        "plt.title(\"Distribuzione individui sposati e non\");"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AfHbNKxkKmKd"
      },
      "source": [
        "Inoltre come viene citato nel seguente articolo https://www.reuters.com/article/us-health-marriage-idUSKBN1JE2XG , \"People who are married may be less likely to develop cardiovascular disease or die from a heart attack or stroke than individuals who aren't, a research review suggests\".\n",
        "\n",
        "Questo suggerisce che lo sbilanciamento dei dati di questa feature sia appunto potenzialmente dannoso per l'accuratezza del nostro modello. Conviene quindi eliminarla dal Dataframe:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "72u1zgD0LYyb"
      },
      "source": [
        "data.drop(columns=[\"ever_married\"], inplace=True)"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SCYo4TTsCnsJ"
      },
      "source": [
        "Elaboriamo adesso i dati inerenti il BMI (o *Body Mass Index*):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "id": "nJLdRA8Ndgxu",
        "outputId": "9daf9f49-2f07-458c-8187-67c9edbd30be"
      },
      "source": [
        "data.bmi.plot.hist(bins=40, color=\"darkorange\");\n",
        "plt.title(\"Distribuzione BMI negli individui sotto osservazione\")\n",
        "plt.xlabel(\"BMI\");"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I6kmZWM5exi3"
      },
      "source": [
        "L'indice di massa corporea come si nota presenta una deviazione standard molto bassa in quanto la maggior parte degli individui presenta valori nell'intervallo tra `20` e `40`, dove `18.5-24,99` è la fascia in cui rientrano gli individui normopeso, mentre oltre questa fascia vi sono gli individui sovrappeso e obesi."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xy2i80Af32Nv"
      },
      "source": [
        "Rappresentazione tramite barplot di `Residence_Type` \n",
        "a confronto con `stroke`"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "id": "B8wHrQWX4I5K",
        "outputId": "be5a6721-b3f4-46a3-ee1f-83353fa679d8"
      },
      "source": [
        "sns.set_theme(style=\"whitegrid\")\n",
        "plt.title(\"Correlazione tra tipo di residenza e incidenza di ictus\")\n",
        "sns.barplot(x=\"Residence_type\", y=\"stroke\", data=data);"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "h77pkA5g5DXH"
      },
      "source": [
        "Possiamo notare come i soggetti con residenza urbana siano statisticamente leggermente più soggetti alla possibilità di colpi apoplettici. Appare quindi che questa feature non sia indicativa nel rilevare la possibilità di ictus.\n",
        "\n",
        "Mostriamo ora la correlazione che `hypertension` e `heart_disease` hanno con l'incidenza di ictus e come si correlano tra di loro:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 338
        },
        "id": "g4Y_t5sqz065",
        "outputId": "641e90a6-c8fe-4e0d-a452-55cd67458d5b"
      },
      "source": [
        "plt.figure(figsize=(10, 5))\n",
        "\n",
        "sns.lineplot(x=\"heart_disease\", y=\"stroke\", label=\"heart_disease\", data=data);\n",
        "plt.legend()\n",
        "\n",
        "sns.lineplot(x=\"hypertension\", y=\"stroke\", label=\"hypertension\", data=data);\n",
        "plt.legend()\n",
        "plt.title(\"Incidenza di ictus in individui con ipertensione e/o malattie cardiache\");\n",
        "plt.xlabel(\"\");"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QgrzAAjo8gGh"
      },
      "source": [
        "Dal grafico emerge ovviamente che queste sintomatologie influenzano la possibilità di colpi apoplettici (seppur non fortemente). Si palesa anche una certa correlazione tra le due feature `hypertension` e `heart_disease`. Esploreremo meglio questa correlazione in seguito usando la **regolarizzazione L1**.\n",
        "\n",
        "Esploriamo la correlazione che invece con il tipo di lavoro che gli individui svolgono:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        },
        "id": "RovlmRvLC9tq",
        "outputId": "160b4489-06db-432b-d9e9-a7eaf300dc07"
      },
      "source": [
        "sns.barplot(x=\"work_type\", y=\"stroke\", data=data);"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kB4WaB70DtGA"
      },
      "source": [
        "Gli individui che non hanno mai lavorato si rivelano, sorprendentemente, i meno propensi a colpi apoplettici, addirittura con meno episodi rispetto ai bambini. Probabilmente questa feature avrà una forte correlazione inversa con quella di ictus."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dStp5c6A7F-s"
      },
      "source": [
        "Ora rappresentiamo su barplot la correlazione tra ictus e stato di fumatore:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "id": "IftyPsgg5840",
        "outputId": "f8c8693a-4939-45e6-aafe-364d3710f2dc"
      },
      "source": [
        "bars = sns.barplot(x=\"smoking_status\", y=\"stroke\", data=data)\n",
        "plt.title(\"Correlazione tra fumo e incidenza di ictus\")\n",
        "bars.set_xticklabels(bars.get_xticklabels(), fontsize=10);"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yLWGov6WqGQ3"
      },
      "source": [
        "Dal grafico notiamo che gli individui che in passato erano fumatori hanno un'incidenza di colpi apoplettici di gran lunga maggiore rispetto a quella delle altre categorie in questione. Possiamo notare anche la presenza della colonna con valore `Unknown` che in questo caso risulta problematica nel corretto sviluppo del modello. Ce ne occuperemo nella fase di preparazione dei dati."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hhxX3SOj4G7L"
      },
      "source": [
        "Approfondendo lo studio dei dati, plottiamo la matrice di correlazione delle feature numeriche del Dataframe:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 358
        },
        "id": "G3h9RjeM1lzV",
        "outputId": "7788825f-7d95-408e-99f3-1478d5b7c9b4"
      },
      "source": [
        "corr = data.corr()\n",
        "mask = np.triu(np.ones_like(corr, dtype=bool))\n",
        "sns.heatmap(corr, mask=mask, annot=True, cmap=\"Reds\");"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NQigMGQP4O-P"
      },
      "source": [
        "Notiamo come `age` sia molto correlata con tutte le feature e come tra le altre feature le correlazioni siano molto basse. D'altra parte `stroke` risulta influenzata in primis da `age` e di pari passo da `hypertension`, `heart_disease` e `avg_glucose_level` . Approfondiremo lo studio su quale feature siano più rilevanti in seguito. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7LX4xHkiAiII"
      },
      "source": [
        "##Preparazione dei dati"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hInOIDeEzOlU"
      },
      "source": [
        "###Elaborazione valori nulli"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ub2Z-ldbiKG8"
      },
      "source": [
        "Andiamo ora a elaborare i dati presenti nel Dataframe affinchè vengano individuate le feature rilevanti ed ignorate quelle irrilevanti."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4AHcW7IGN_WZ"
      },
      "source": [
        "Iniziamo andando a visualizzare le righe del Dataframe che contengono valori nulli nelle colonne di tipo numerico:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WjGkZe9GO0-4",
        "outputId": "0b01d6d0-3add-4979-c800-a3b3c5b2d739"
      },
      "source": [
        "data.shape[0]-data.count()"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "gender                 0\n",
              "age                    0\n",
              "hypertension           0\n",
              "heart_disease          0\n",
              "work_type              0\n",
              "Residence_type         0\n",
              "avg_glucose_level      0\n",
              "bmi                  201\n",
              "smoking_status         0\n",
              "stroke                 0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G65_6CFTtKnr"
      },
      "source": [
        "Notiamo la presenza di 201 righe con valore di `bmi` nullo. Procediamo rimuovendo queste righe:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FB-2X6kztff3",
        "outputId": "427a0406-62a2-488e-d892-dc72c6d4041d"
      },
      "source": [
        "data.dropna(inplace=True)\n",
        "data.shape[0]-data.count()"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "gender               0\n",
              "age                  0\n",
              "hypertension         0\n",
              "heart_disease        0\n",
              "work_type            0\n",
              "Residence_type       0\n",
              "avg_glucose_level    0\n",
              "bmi                  0\n",
              "smoking_status       0\n",
              "stroke               0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MMdsf1Uctrta"
      },
      "source": [
        "Nei dati della colonna `smoking_status` abbiamo notato la presenza dei valori `Unknown`. Andiamo a visualizzare effettivamente quanti dati presentano questo valore:\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "rbUWwzBAuTbV",
        "outputId": "dd557070-e035-43bf-9225-40c229abfa1b"
      },
      "source": [
        "data.smoking_status.value_counts().plot.pie(autopct=\"%.1f%%\");\n",
        "plt.title(\"Distribuzione individui fumatori e non\");\n",
        "plt.ylabel(\"\");"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Z5fqYrXPuf0j"
      },
      "source": [
        "La colonna `Unknown` copre il `30.2%` delle istanze ed essendo questo valore non deducibile da altre colonne, conviene eliminare le righe con questo valore."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tS-ADix4wlwe"
      },
      "source": [
        "data = data[data.smoking_status != \"Unknown\"]"
      ],
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "3f-OpvVfw-eV",
        "outputId": "7706487a-0b1e-4c52-ad8d-4e57d855cc96"
      },
      "source": [
        "data.smoking_status.value_counts().plot.pie(autopct=\"%.1f%%\");\n",
        "plt.title(\"Distribuzione individui fumatori e non\");\n",
        "plt.ylabel(\"\");"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RaxUJRhpzTFf"
      },
      "source": [
        "###Binarizzazione"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Eqb07ROAzXA6"
      },
      "source": [
        "Attualmente nel Dataframe attuale ci sono 4 variabili categoriche:\n",
        "-   `gender`\n",
        "-   `work_type`\n",
        "- `Residence_type`\n",
        "- `smoking_status`\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fU2ktAaS3GfO"
      },
      "source": [
        "Affinchè esse siano utilizzabili nel nostro modello è necessario che vengano mappate usando delle variabili binarie. Usiamo la funzione di binarizzazione \n",
        "`get_dummies()`:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PTAHoPNA49Xn"
      },
      "source": [
        "data = pd.get_dummies(data, columns=[\"gender\", \"work_type\", \"Residence_type\", \"smoking_status\"], prefix=[\"gender\", \"work\", \"Residence\", \"smoking\"])"
      ],
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "BfxrQb1R6ln6",
        "outputId": "2898c0f4-cfd0-45ad-fd51-30ac5d7caa06"
      },
      "source": [
        "data.head()"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>age</th>\n",
              "      <th>hypertension</th>\n",
              "      <th>heart_disease</th>\n",
              "      <th>avg_glucose_level</th>\n",
              "      <th>bmi</th>\n",
              "      <th>stroke</th>\n",
              "      <th>gender_Female</th>\n",
              "      <th>gender_Male</th>\n",
              "      <th>gender_Other</th>\n",
              "      <th>work_Govt_job</th>\n",
              "      <th>work_Never_worked</th>\n",
              "      <th>work_Private</th>\n",
              "      <th>work_Self-employed</th>\n",
              "      <th>work_children</th>\n",
              "      <th>Residence_Rural</th>\n",
              "      <th>Residence_Urban</th>\n",
              "      <th>smoking_Unknown</th>\n",
              "      <th>smoking_formerly smoked</th>\n",
              "      <th>smoking_never smoked</th>\n",
              "      <th>smoking_smokes</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>67.0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>228.69</td>\n",
              "      <td>36.6</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>80.0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>105.92</td>\n",
              "      <td>32.5</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>49.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>171.23</td>\n",
              "      <td>34.4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>79.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>174.12</td>\n",
              "      <td>24.0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>81.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>186.21</td>\n",
              "      <td>29.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    age  hypertension  ...  smoking_never smoked  smoking_smokes\n",
              "0  67.0             0  ...                     0               0\n",
              "2  80.0             0  ...                     1               0\n",
              "3  49.0             0  ...                     0               1\n",
              "4  79.0             1  ...                     1               0\n",
              "5  81.0             0  ...                     0               0\n",
              "\n",
              "[5 rows x 20 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WNrLijfX6BlZ"
      },
      "source": [
        "Procediamo rimuovendo le colonne associate ai valori non più presenti all'interno del Dataframe (`Other` e `Unknown`): "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R0ZEyKH66SCt"
      },
      "source": [
        "data.drop(columns=[\"gender_Other\", \"smoking_Unknown\"], inplace=True)"
      ],
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NMjSYxHX8oCl"
      },
      "source": [
        "Il modello si presenta ora pronto alla fase successiva di preparazione."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6ZVlkx3H8uSU"
      },
      "source": [
        "##Preparazione del modello"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9A3VyoXEB5DX"
      },
      "source": [
        "###Standardizzazione, regolarizzazione e studio feature più importanti\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6x5xed9yCkDv"
      },
      "source": [
        "Iniziamo importando le librerie necessarie:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A2Eo6GNdwq-w"
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from sklearn.linear_model import Perceptron\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.preprocessing import StandardScaler"
      ],
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "s2Dfu0B6w-WM"
      },
      "source": [
        "Dopodichè estraiamo dal Dataframe la variabile da prevedere (`stroke`) e il Dataframe costituito unicamente dalle altre variabili:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iqfQGT0bxEiZ"
      },
      "source": [
        "X = data.drop(columns=\"stroke\")\n",
        "y = data[\"stroke\"]"
      ],
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "asl1fd0hxe_J"
      },
      "source": [
        "Procediamo dunque suddividendo i dati in training e validation set tramite metodo **holdout**:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t2s2a6FTyKlM"
      },
      "source": [
        "X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=1/3, random_state=42)"
      ],
      "execution_count": 39,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jg_o7pNDyuo5"
      },
      "source": [
        "Eseguiamo adesso una prova usando i set appena creati per valutare un semplice modello basato su Perceptron:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AUsLoSXOzGO-"
      },
      "source": [
        "model = Pipeline([\n",
        "                  (\"perc\", Perceptron(n_jobs=-1, random_state=42))\n",
        "])"
      ],
      "execution_count": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_hxhFmVu3ozG",
        "outputId": "390053e0-90f1-45ba-f005-9c76f3ebb717"
      },
      "source": [
        "model.fit(X_train, y_train)\n",
        "print(\"R-squared coefficient:\")\n",
        "model.score(X_val, y_val)"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "R-squared coefficient:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9159369527145359"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kCh35wYXG0Ex"
      },
      "source": [
        "Il nostro modello, nonostante le variabili non siano normalizzate, produce già ottimi risultati con un tasso di accuratezza che arriva a `91,6 %` di accuratezza. Ciò accade perchè le variabili presenti non presentano ordini di grandezza troppo distanti.\n",
        "\n",
        "Proviamo adesso a **standardizzare** le feature e vedere come ciò influisca sui nostri risultati:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WpCQl_KrH_w8"
      },
      "source": [
        "std_model = Pipeline([\n",
        "                      (\"scaler\", StandardScaler()),\n",
        "                      (\"perc\", Perceptron(n_jobs=-1, random_state=42))\n",
        "])"
      ],
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3XC-u56UITe4",
        "outputId": "b8f1e2ba-5851-4624-d767-90d4863421eb"
      },
      "source": [
        "std_model.fit(X_train, y_train)\n",
        "print(\"R-squared coefficient:\")\n",
        "std_model.score(X_val, y_val)"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "R-squared coefficient:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9369527145359019"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d8b1hSZCJn_a"
      },
      "source": [
        "Appare che standardizzando le feature il tasso di accuratezza, seppur di poco, salga ancora.\n",
        "Proviamo adesso ad aggiungere una **regolarizzazione con norma L1** per individuare le feature più rilevanti e eliminare eventuali feature inutili:\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OAIG97GjKLma"
      },
      "source": [
        "std_l1_model = Pipeline([\n",
        "                      (\"scaler\", StandardScaler()),\n",
        "                      (\"perc\", Perceptron(penalty=\"l1\", alpha=0.0001, n_jobs=-1, random_state=42))\n",
        "])"
      ],
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AwKo7oERKgOY",
        "outputId": "d35b0492-a273-443b-a3c0-734c675209ed"
      },
      "source": [
        "std_l1_model.fit(X_train, y_train)\n",
        "print(\"R-squared coefficient:\")\n",
        "std_l1_model.score(X_val, y_val)"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "R-squared coefficient:\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.936077057793345"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "3edXsmTgPdYJ",
        "outputId": "c41f9113-81af-4254-8ca2-375305663249"
      },
      "source": [
        "col_coef = pd.DataFrame(std_l1_model.named_steps[\"perc\"].coef_[0], columns=[\"coefficients\"], index=X.columns)\n",
        "col_coef"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>coefficients</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>age</th>\n",
              "      <td>1.466258</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hypertension</th>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>heart_disease</th>\n",
              "      <td>-1.926350</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>avg_glucose_level</th>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>bmi</th>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>gender_Female</th>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>gender_Male</th>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>work_Govt_job</th>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>work_Never_worked</th>\n",
              "      <td>-14.150164</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>work_Private</th>\n",
              "      <td>0.042962</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>work_Self-employed</th>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>work_children</th>\n",
              "      <td>-1.004935</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Residence_Rural</th>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Residence_Urban</th>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>smoking_formerly smoked</th>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>smoking_never smoked</th>\n",
              "      <td>-1.494250</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>smoking_smokes</th>\n",
              "      <td>1.063614</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                         coefficients\n",
              "age                          1.466258\n",
              "hypertension                 0.000000\n",
              "heart_disease               -1.926350\n",
              "avg_glucose_level            0.000000\n",
              "bmi                          0.000000\n",
              "gender_Female                0.000000\n",
              "gender_Male                  0.000000\n",
              "work_Govt_job                0.000000\n",
              "work_Never_worked          -14.150164\n",
              "work_Private                 0.042962\n",
              "work_Self-employed           0.000000\n",
              "work_children               -1.004935\n",
              "Residence_Rural              0.000000\n",
              "Residence_Urban              0.000000\n",
              "smoking_formerly smoked      0.000000\n",
              "smoking_never smoked        -1.494250\n",
              "smoking_smokes               1.063614"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ASYuwMxGHowJ",
        "outputId": "2f083c17-ac5e-4bb4-9f1e-64a8e4c63a24"
      },
      "source": [
        "col_coef[col_coef.coefficients == 0].size"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "10"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 47
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7q-WcWR5Bycn"
      },
      "source": [
        "Il modello ha una perdità di accuratezza infima e il peso di molte feature si annulla (si azzerano ben 10 feature su 17). Possiamo notare come emerga come feature più importante la colonna `work_Never_worked` che presenta un peso fortemente negativo. Possiamo anche notare come la feature `hypertension` si sia annullata a differenza di `heart_disease`risolvendo la presunta correlazione tra le due feature notata nel paragrafo di analisi dei dati.\n",
        "\n",
        "Procediamo eliminando le feature non rilevanti:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f9t91EBu5xdn"
      },
      "source": [
        "data.drop(columns=col_coef[col_coef.coefficients == 0].index, inplace=True)"
      ],
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6MZICiyI5t_3"
      },
      "source": [
        "X = data.drop(columns=\"stroke\")\n",
        "y = data[\"stroke\"]"
      ],
      "execution_count": 49,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0eBAbltnH2g2"
      },
      "source": [
        "E' importante notare come gli score di accuratezza sopra ottenuti non siano da ritenersi attendibili in quanto il dataset in uso presenta classi fortemente sbilanciate, comportando la generazione di un modello molto sbilanciato (risolveremo questa problematica nel paragrafo di modellazione), come si può notare dalla seguente matrice di confusione:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "v1mJIEQaIltP"
      },
      "source": [
        "from sklearn.metrics import confusion_matrix"
      ],
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "er2LJlqRIs6Q",
        "outputId": "fc67de67-86e4-4bb2-fd8b-1dd0d01ce4e2"
      },
      "source": [
        "index = [\"no stroke\", \"stroke\"]\n",
        "classes = [\"no stroke[P]\", \"stroke[P]\"]\n",
        "pd.DataFrame(confusion_matrix(y_val, std_l1_model.predict(X_val)), columns=classes, index = index)"
      ],
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>no stroke[P]</th>\n",
              "      <th>stroke[P]</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>no stroke</th>\n",
              "      <td>1068</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>stroke</th>\n",
              "      <td>62</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "           no stroke[P]  stroke[P]\n",
              "no stroke          1068         11\n",
              "stroke               62          1"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-vSuWMJcNJgz"
      },
      "source": [
        "##Modellazione\n",
        "Partiamo importando le librerie che ci serviranno per la fase di modellazione:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F4xY14TYR-gk"
      },
      "source": [
        "import warnings\n",
        "warnings.filterwarnings('ignore')\n",
        "\n",
        "from sklearn.metrics import f1_score\n",
        "from sklearn.model_selection import KFold\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from imblearn.over_sampling import BorderlineSMOTE\n",
        "from imblearn.over_sampling import SMOTE\n",
        "from sklearn.svm import LinearSVC\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.metrics import recall_score\n",
        "from sklearn.metrics import precision_score\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "import math\n",
        "from sklearn import tree\n",
        "import graphviz"
      ],
      "execution_count": 52,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lPbNw5p5G7HD"
      },
      "source": [
        "Dato il forte sbilanciamento delle classi, è necessario ricorrere a una strategia di **oversampling** per fare in modo che il nostro modello non risulti troppo sbilanciato. In questo caso ricorriamo a **BorderlineSMOTE**, scelto dopo un testing fatto su SMOTE applicato a tutti i modelli sotto elencati. Altre tecniche in grado di risolvere tale problematica non verranno trattate nel seguente documento. Procediamo rinnovando training e validation set e facendo oversampling per compensare lo sbilanciamento:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PJffh-z_48dj"
      },
      "source": [
        "X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=1/3, random_state=42)"
      ],
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fYvhV7PySof1"
      },
      "source": [
        "sm = BorderlineSMOTE(random_state=42)\n",
        "X_res, y_res = sm.fit_resample(X, y)\n",
        "X_train_res, y_train_res = sm.fit_resample(X_train, y_train)\n",
        "X_train_res = pd.DataFrame(X_train_res, columns=X_train.columns)"
      ],
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "0Z591rnDjHPz",
        "outputId": "4a2d808a-60e0-4ca7-d4ab-f7479a302640"
      },
      "source": [
        "X_train_res.head()"
      ],
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>age</th>\n",
              "      <th>heart_disease</th>\n",
              "      <th>work_Never_worked</th>\n",
              "      <th>work_Private</th>\n",
              "      <th>work_children</th>\n",
              "      <th>smoking_never smoked</th>\n",
              "      <th>smoking_smokes</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>31.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>64.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>30.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>36.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>52.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    age  heart_disease  ...  smoking_never smoked  smoking_smokes\n",
              "0  31.0            0.0  ...                   1.0             0.0\n",
              "1  64.0            0.0  ...                   1.0             0.0\n",
              "2  30.0            0.0  ...                   1.0             0.0\n",
              "3  36.0            0.0  ...                   0.0             1.0\n",
              "4  52.0            0.0  ...                   1.0             0.0\n",
              "\n",
              "[5 rows x 7 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 55
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AwppIATQUZDI",
        "outputId": "0457485a-6c4e-4aac-abb0-0728ce59560f"
      },
      "source": [
        "y_train_res.sum()"
      ],
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2166"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0TTr0KI8bGku"
      },
      "source": [
        "Creaiamo dei dizionari per memorizzare i valori che otteremo dai modelli analizzati:\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1OT6uDlmbXGw"
      },
      "source": [
        "scores = {}\n",
        "f1_scores = {}\n",
        "precision = {}\n",
        "recall = {}"
      ],
      "execution_count": 57,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DcELXuAIl8FM"
      },
      "source": [
        "###Perceptron"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x3eU6SzvmAdN"
      },
      "source": [
        "Iniziamo la fase di modellazione provando un modello basato su Perceptron, questa volta usando GridSearch per trovare i migliori parametri possibili. Analizziamo i seguenti iperparametri:\n",
        "- standardizzazione o meno delle feature\n",
        "- tipo di regolarizzazione del modello\n",
        "- peso della regolarizzazione\n",
        "- stima o meno dell'intercetta"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h11ZVERqf_2x"
      },
      "source": [
        "kfold = KFold(n_splits=5, shuffle=True, random_state=42)"
      ],
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tvHgxx3s3VSI"
      },
      "source": [
        "perc_model = Pipeline([\n",
        "    (\"scaler\", StandardScaler()),\n",
        "    (\"perc\", Perceptron(n_jobs=-1, random_state=42))\n",
        "])\n",
        "\n",
        "#print(per_model.get_params())\n",
        "\n",
        "perc_grid = {\n",
        "    \"scaler\": [None, StandardScaler()],\n",
        "    \"perc__penalty\": [\"l2\", \"l1\", \"elasticnet\"],\n",
        "    \"perc__alpha\": np.logspace(-3, 3, 10),\n",
        "    \"perc__fit_intercept\": [False, True]\n",
        "}"
      ],
      "execution_count": 59,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3lPEZAyd3wnY",
        "outputId": "5e0ba5ea-981f-435b-cec9-c06315c7c14a"
      },
      "source": [
        "%%time\n",
        "\n",
        "perc_gv = GridSearchCV(perc_model, perc_grid, cv=kfold, n_jobs=-1)\n",
        "perc_gv.fit(X_train_res, y_train_res)"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 1.26 s, sys: 67.8 ms, total: 1.33 s\n",
            "Wall time: 6.51 s\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2VaB81Zg4Nia",
        "outputId": "9dcec51d-076a-4221-db88-a98e6dbe4de7"
      },
      "source": [
        "print(\"Punteggio migliore: {score}\".format(score=perc_gv.score(X_val, y_val)));\n",
        "print(\"F1 score: {score}\".format(score=f1_score(y_val, perc_gv.predict(X_val), average=\"binary\")));\n",
        "print(\"Precision score: {score}\".format(score=precision_score(y_val, perc_gv.predict(X_val))))\n",
        "print(\"Recall score: {score}\".format(score=recall_score(y_val, perc_gv.predict(X_val))))\n",
        "print(\"Parametri migliori: {params}\".format(params=perc_gv.best_params_))\n",
        "\n",
        "scores[\"perc\"] = perc_gv.score(X_val, y_val);\n",
        "f1_scores[\"perc\"] = f1_score(y_val, perc_gv.predict(X_val), average=\"binary\")\n",
        "precision[\"perc\"] = precision_score(y_val, perc_gv.predict(X_val))\n",
        "recall[\"perc\"] = recall_score(y_val, perc_gv.predict(X_val))"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Punteggio migliore: 0.7110332749562172\n",
            "F1 score: 0.23255813953488372\n",
            "Precision score: 0.1362397820163488\n",
            "Recall score: 0.7936507936507936\n",
            "Parametri migliori: {'perc__alpha': 0.001, 'perc__fit_intercept': False, 'perc__penalty': 'l1', 'scaler': StandardScaler(copy=True, with_mean=True, with_std=True)}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "id": "MfPLsMR7YqFq",
        "outputId": "b6f40534-50ed-4ea4-860a-221b71a0d141"
      },
      "source": [
        "index = [\"no stroke\", \"stroke\"]\n",
        "classes = [\"no stroke[P]\", \"stroke[P]\"]\n",
        "perc_conf = pd.DataFrame(confusion_matrix(y_res, perc_gv.predict(X_res)), columns=classes, index = index)\n",
        "perc_conf"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>no stroke[P]</th>\n",
              "      <th>stroke[P]</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>no stroke</th>\n",
              "      <td>2268</td>\n",
              "      <td>977</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>stroke</th>\n",
              "      <td>572</td>\n",
              "      <td>2673</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "           no stroke[P]  stroke[P]\n",
              "no stroke          2268        977\n",
              "stroke              572       2673"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Re78pKudLo38"
      },
      "source": [
        "Come era prevedibile l'uso di SMOTE ha comportato una significativa riduzione del tasso di accuratezza. Dalla matrice di confusione si nota che il modello appare ancora piuttosto sbilanciato. Procediamo a migliorarlo usando la **Regressione Logistica**."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ryEiXlKWjNXy"
      },
      "source": [
        "###Regressione Logistica"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Z7QYeN0FeDAJ"
      },
      "source": [
        "Iniziamo la fase di modellazione provando un modello di Regressione Logistica con scelti dei migliori parametri tramite uso di GridSearch, con test sui seguenti parametri:\n",
        "- standardizzazione o meno delle feature\n",
        "- tipo di regolarizzazione del modello"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yFK6LbqWeaNh"
      },
      "source": [
        "log_model = Pipeline([\n",
        "    (\"scaler\", StandardScaler()),\n",
        "    (\"log_reg\", LogisticRegression(solver='saga', random_state=42))\n",
        "])\n",
        "\n",
        "log_grid = {\n",
        "    \"scaler\": [None, StandardScaler()],\n",
        "    \"log_reg__penalty\": [\"l2\", \"l1\"],\n",
        "    \"log_reg__C\": np.logspace(-3, 3, 10),\n",
        "    \"log_reg__fit_intercept\": [False, True]\n",
        "}"
      ],
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "A3ILogBY31KZ",
        "outputId": "2fe81d76-c293-453b-ded6-f30b53bd2e91"
      },
      "source": [
        "%%time\n",
        "\n",
        "log_gv = GridSearchCV(log_model, log_grid, cv=kfold, n_jobs=-1)\n",
        "log_gv.fit(X_train_res, y_train_res)"
      ],
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 931 ms, sys: 83.6 ms, total: 1.01 s\n",
            "Wall time: 17.1 s\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ylNFbrrHg_pa",
        "outputId": "0c9e7fc3-e4d4-480d-caaa-525fb4758ab4"
      },
      "source": [
        "print(\"Punteggio migliore: {score}\".format(score=log_gv.score(X_val, y_val)));\n",
        "print(\"F1 score: {score}\".format(score=f1_score(y_val, log_gv.predict(X_val), average=\"binary\")));\n",
        "print(\"Precision score: {score}\".format(score=precision_score(y_val, log_gv.predict(X_val))))\n",
        "print(\"Recall score: {score}\".format(score=recall_score(y_val, log_gv.predict(X_val))))\n",
        "print(\"Parametri migliori: {params}\".format(params=log_gv.best_params_))\n",
        "\n",
        "scores[\"log\"] = log_gv.score(X_val, y_val);\n",
        "f1_scores[\"log\"] = f1_score(y_val, log_gv.predict(X_val), average=\"binary\")\n",
        "precision[\"log\"] = precision_score(y_val, log_gv.predict(X_val))\n",
        "recall[\"log\"] = recall_score(y_val, log_gv.predict(X_val))"
      ],
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Punteggio migliore: 0.7171628721541156\n",
            "F1 score: 0.2327790973871734\n",
            "Precision score: 0.13687150837988826\n",
            "Recall score: 0.7777777777777778\n",
            "Parametri migliori: {'log_reg__C': 0.021544346900318832, 'log_reg__fit_intercept': True, 'log_reg__penalty': 'l1', 'scaler': StandardScaler(copy=True, with_mean=True, with_std=True)}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "toU65tz5aW-p",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "outputId": "ee50db6d-d015-468d-edfc-b22ff00800b4"
      },
      "source": [
        "log_conf = pd.DataFrame(confusion_matrix(y_res, log_gv.predict(X_res)), columns=classes, index = index)\n",
        "log_conf"
      ],
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>no stroke[P]</th>\n",
              "      <th>stroke[P]</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>no stroke</th>\n",
              "      <td>2333</td>\n",
              "      <td>912</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>stroke</th>\n",
              "      <td>475</td>\n",
              "      <td>2770</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "           no stroke[P]  stroke[P]\n",
              "no stroke          2333        912\n",
              "stroke              475       2770"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 66
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hnXrGz6IMOU1"
      },
      "source": [
        "Il tasso di accuratezza si abbassa ancora ma ciononostante la matrice di confusione indica che abbiamo un netto miglioramento nelle capacità predittiva del modello che adesso risulta notevolmente più affidabile (specialmente nel caso di falsi negativi) rispetto al modello basato su **Perceptron**. D'altra parte il tasso di falsi positivi è ancora molto alto. Proviamo in seguito una modellazione tramite **Support Vector Machines**."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "km1M4TEzNOjP"
      },
      "source": [
        "###SVM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pYKrkLnFOagn"
      },
      "source": [
        "svc_model = Pipeline([\n",
        "    (\"scaler\", StandardScaler()),\n",
        "    (\"svc\", LinearSVC(dual=False, random_state=42))\n",
        "])\n",
        "\n",
        "svc_grid = {\n",
        "    \"scaler\": [None, StandardScaler()],\n",
        "    \"svc__penalty\": [\"l2\", \"l1\"],\n",
        "    \"svc__loss\": [\"hinge\", \"squared_hinge\"],\n",
        "    \"svc__fit_intercept\": [False, True],\n",
        "    \"svc__class_weight\": [None, \"balanced\"],\n",
        "    \"svc__C\": np.logspace(-3, 3, 10)\n",
        "}"
      ],
      "execution_count": 67,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7JD8HVudP9Q7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "32f54c71-b5e6-4214-c7d6-e66fcddf1d12"
      },
      "source": [
        "%%time\n",
        "\n",
        "svc_gv = GridSearchCV(svc_model, svc_grid, cv=kfold, n_jobs=-1)\n",
        "svc_gv.fit(X_train_res, y_train_res)"
      ],
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 2.38 s, sys: 80.2 ms, total: 2.46 s\n",
            "Wall time: 10.4 s\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mKoxV-_zQjOq",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e01c4b33-6a4f-4b9a-9f01-6bd312bcfbe3"
      },
      "source": [
        "print(\"Punteggio migliore: {score}\".format(score=svc_gv.score(X_val, y_val)));\n",
        "print(\"F1 score: {score}\".format(score=f1_score(y_val, svc_gv.predict(X_val), average=\"binary\")));\n",
        "print(\"Precision score: {score}\".format(score=precision_score(y_val, svc_gv.predict(X_val))))\n",
        "print(\"Recall score: {score}\".format(score=recall_score(y_val, svc_gv.predict(X_val))))\n",
        "print(\"Parametri migliori: {params}\".format(params=svc_gv.best_params_))\n",
        "\n",
        "scores[\"svc\"] = svc_gv.score(X_val, y_val);\n",
        "f1_scores[\"svc\"] = f1_score(y_val, svc_gv.predict(X_val), average=\"binary\")\n",
        "precision[\"svc\"] = precision_score(y_val, svc_gv.predict(X_val))\n",
        "recall[\"svc\"] = recall_score(y_val, svc_gv.predict(X_val))"
      ],
      "execution_count": 69,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Punteggio migliore: 0.7092819614711033\n",
            "F1 score: 0.22790697674418606\n",
            "Precision score: 0.1335149863760218\n",
            "Recall score: 0.7777777777777778\n",
            "Parametri migliori: {'scaler': None, 'svc__C': 0.46415888336127775, 'svc__class_weight': 'balanced', 'svc__fit_intercept': True, 'svc__loss': 'squared_hinge', 'svc__penalty': 'l1'}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0YJ3WEhnQnGJ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "outputId": "f1259926-1f25-49dd-8608-f2c998f77e57"
      },
      "source": [
        "svc_conf = pd.DataFrame(confusion_matrix(y_res, svc_gv.predict(X_res)), columns=classes, index = index)\n",
        "svc_conf"
      ],
      "execution_count": 70,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>no stroke[P]</th>\n",
              "      <th>stroke[P]</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>no stroke</th>\n",
              "      <td>2303</td>\n",
              "      <td>942</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>stroke</th>\n",
              "      <td>480</td>\n",
              "      <td>2765</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "           no stroke[P]  stroke[P]\n",
              "no stroke          2303        942\n",
              "stroke              480       2765"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 70
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8aB1SjF_d3Fm"
      },
      "source": [
        "###Decision Tree"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IgPBII75emCn"
      },
      "source": [
        "Usiamo adesso un modello basato su **Albero di Regressione**, strutturato nel seguente modo:\n",
        "- ogni nodo contiene un predicato che a sua volta divide le istanze in input in 2 sottoinsiemi e cosi via ricorsivamente\n",
        "- ogni nodo foglia rappresenta la classificazione attuata sulle istanze che hanno soddisfatto tutti i predicati precedenti\n",
        "\n",
        "Procediamo all'addestramento:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vh5JtbYPALj7"
      },
      "source": [
        "num_features = X.columns.size\n",
        "\n",
        "tree_model = Pipeline([\n",
        "    (\"scaler\", StandardScaler()),\n",
        "    (\"tree\", DecisionTreeClassifier(class_weight=\"balanced\", random_state=42))\n",
        "])\n",
        "\n",
        "#print(tree_model.get_params())\n",
        "\n",
        "tree_grid = {'scaler': [None, StandardScaler()],\n",
        "             'tree__min_samples_split': range(2, 6),\n",
        "             'tree__min_samples_leaf': range(1, 6), \n",
        "             'tree__max_depth': range(2,6),\n",
        "             'tree__max_features': range(2, num_features)}"
      ],
      "execution_count": 71,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BlIikmcYCDzV",
        "outputId": "dd75eb2a-84a8-4e56-edcb-7a1d4be13e72"
      },
      "source": [
        "%%time\n",
        "\n",
        "tree_gv = GridSearchCV(tree_model, tree_grid, cv=kfold, n_jobs=-1)\n",
        "tree_gv.fit(X_train_res, y_train_res)"
      ],
      "execution_count": 72,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 6.68 s, sys: 166 ms, total: 6.85 s\n",
            "Wall time: 25.4 s\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "E0M94sWiCLxE",
        "outputId": "11487ddf-cb5b-40c1-d9f7-a44974be2cb2"
      },
      "source": [
        "print(\"Punteggio migliore: {score}\".format(score=tree_gv.score(X_val, y_val)));\n",
        "print(\"F1 score: {score}\".format(score=f1_score(y_val, tree_gv.predict(X_val), average=\"binary\")));\n",
        "print(\"Precision score: {score}\".format(score=precision_score(y_val, tree_gv.predict(X_val))))\n",
        "print(\"Recall score: {score}\".format(score=recall_score(y_val, tree_gv.predict(X_val))))\n",
        "print(\"Parametri migliori: {params}\".format(params=tree_gv.best_params_))\n",
        "\n",
        "scores[\"tree\"] = tree_gv.score(X_val, y_val);\n",
        "f1_scores[\"tree\"] = f1_score(y_val, tree_gv.predict(X_val), average=\"binary\")\n",
        "precision[\"tree\"] = precision_score(y_val, tree_gv.predict(X_val))\n",
        "recall[\"tree\"] = recall_score(y_val, tree_gv.predict(X_val))"
      ],
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Punteggio migliore: 0.8966725043782837\n",
            "F1 score: 0.20270270270270274\n",
            "Precision score: 0.17647058823529413\n",
            "Recall score: 0.23809523809523808\n",
            "Parametri migliori: {'scaler': None, 'tree__max_depth': 5, 'tree__max_features': 3, 'tree__min_samples_leaf': 1, 'tree__min_samples_split': 2}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "j79t4BrBhLkO"
      },
      "source": [
        "Il modello ottiene i valori migliori rispetto ai modelli precedenti con un tasso di accuratezza di gran lunga maggiore come possiamo notare dalla seguente tabella di confronto:"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qrLElFseo0XS"
      },
      "source": [
        "Tuttavia per ottenere una stima valida, osserviamo la matrice di confusione di questo modello:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "id": "U-ZgOfbcodqP",
        "outputId": "d4d79a25-18e4-45ce-9eae-4e5d355289f5"
      },
      "source": [
        "tree_conf = pd.DataFrame(confusion_matrix(y_res, tree_gv.predict(X_res)), columns=classes, index = index)\n",
        "tree_conf"
      ],
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>no stroke[P]</th>\n",
              "      <th>stroke[P]</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>no stroke</th>\n",
              "      <td>3049</td>\n",
              "      <td>196</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>stroke</th>\n",
              "      <td>882</td>\n",
              "      <td>2363</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "           no stroke[P]  stroke[P]\n",
              "no stroke          3049        196\n",
              "stroke              882       2363"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 74
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "175Xlbk7pZcp"
      },
      "source": [
        "Il modello, nonostante l'uso di BorderlineSmote, risulta fortemente sbilanciato e presenta un forte numero di falsi negativi. Questo implica che in questo caso i valori di coefficiente di correlazione e F1 Score non siano attendibili nel descrivere l'affidabilità del modello e che il modello sia probabilmente in **overfitting**."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xku6wJ-ThoJb"
      },
      "source": [
        "### Random Forest"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G6WFECAlWwu3"
      },
      "source": [
        "forest_model = Pipeline([\n",
        "    (\"scaler\", StandardScaler()),\n",
        "    (\"forest\", RandomForestClassifier(n_jobs=-1, random_state=42))\n",
        "])\n",
        "\n",
        "forest_grid = {\"scaler\": [None, StandardScaler()],\n",
        "             'forest__n_estimators': range(5, 10),\n",
        "             'forest__min_samples_split': range(2, 5), \n",
        "             'forest__max_depth': [None] + [i for i in range(1, 3)],\n",
        "             'forest__max_features': [int(math.sqrt(num_features)), num_features - 1]}"
      ],
      "execution_count": 75,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qDNj80hDaY3B",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8b187cc7-d0c4-4b5f-de3e-ff3d5eadea39"
      },
      "source": [
        "%%time\n",
        "\n",
        "forest_gv = GridSearchCV(forest_model, forest_grid, cv=kfold, n_jobs=-1)\n",
        "forest_gv.fit(X_train_res, y_train_res)"
      ],
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 5.47 s, sys: 324 ms, total: 5.8 s\n",
            "Wall time: 1min 37s\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e8V0tqgjalZy",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7b28f60f-85d9-441b-ab2d-2ebad2145e01"
      },
      "source": [
        "print(\"Punteggio migliore: {score}\".format(score=forest_gv.score(X_val, y_val)));\n",
        "print(\"Precision score: {score}\".format(score=precision_score(y_val, forest_gv.predict(X_val))))\n",
        "print(\"Recall score: {score}\".format(score=recall_score(y_val, forest_gv.predict(X_val))))\n",
        "print(\"Parametri migliori: {params}\".format(params=forest_gv.best_params_))\n",
        "\n",
        "scores[\"forest\"] = forest_gv.score(X_val, y_val);\n",
        "f1_scores[\"forest\"] = f1_score(y_val, forest_gv.predict(X_val), average=\"binary\")\n",
        "precision[\"forest\"] = precision_score(y_val, forest_gv.predict(X_val))\n",
        "recall[\"forest\"] = recall_score(y_val, forest_gv.predict(X_val))"
      ],
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Punteggio migliore: 0.8922942206654991\n",
            "Precision score: 0.13414634146341464\n",
            "Recall score: 0.1746031746031746\n",
            "Parametri migliori: {'forest__max_depth': None, 'forest__max_features': 2, 'forest__min_samples_split': 3, 'forest__n_estimators': 5, 'scaler': None}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YsUZ9cwqa7CD",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "outputId": "a3b6e7b1-1f2c-4bcf-9987-29bf74a1118b"
      },
      "source": [
        "forest_conf = pd.DataFrame(confusion_matrix(y_res, forest_gv.predict(X_res)), columns=classes, index = index)\n",
        "forest_conf"
      ],
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>no stroke[P]</th>\n",
              "      <th>stroke[P]</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>no stroke</th>\n",
              "      <td>3075</td>\n",
              "      <td>170</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>stroke</th>\n",
              "      <td>506</td>\n",
              "      <td>2739</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "           no stroke[P]  stroke[P]\n",
              "no stroke          3075        170\n",
              "stroke              506       2739"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 78
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CYz-0J4G2QGm"
      },
      "source": [
        "Il modello `RandomForestClassifier` si comporta infine in maniera molto simile al `DecisionTreeClassifier`, come era prevedibile essendo una Random Forest un insieme di Decision Tree.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ogC379Pujv4i"
      },
      "source": [
        "##Valutazione dei vari modelli utilizzati"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DhcWojrEj1um"
      },
      "source": [
        "Confrontiamo adesso i vari modelli e i valori ottenuti da ciascuno di essi.\n",
        "Analizziamo **coefficiente R^2** e **F1 score**:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1DaZpH_jDPfk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "074ffd75-f682-4b64-ca3e-c47ae0a93cd6"
      },
      "source": [
        "pd.DataFrame.from_dict(scores, orient=\"index\", columns=[\"R^2 score\"])"
      ],
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>R^2 score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>perc</th>\n",
              "      <td>0.711033</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log</th>\n",
              "      <td>0.717163</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>svc</th>\n",
              "      <td>0.709282</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tree</th>\n",
              "      <td>0.896673</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>forest</th>\n",
              "      <td>0.892294</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        R^2 score\n",
              "perc     0.711033\n",
              "log      0.717163\n",
              "svc      0.709282\n",
              "tree     0.896673\n",
              "forest   0.892294"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 79
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WfWAsEJCDPfm",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "58644a9e-6d89-4823-8824-15527976584c"
      },
      "source": [
        "pd.DataFrame.from_dict(f1_scores, orient=\"index\", columns=[\"F1 score\"])"
      ],
      "execution_count": 80,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>F1 score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>perc</th>\n",
              "      <td>0.232558</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log</th>\n",
              "      <td>0.232779</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>svc</th>\n",
              "      <td>0.227907</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tree</th>\n",
              "      <td>0.202703</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>forest</th>\n",
              "      <td>0.151724</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        F1 score\n",
              "perc    0.232558\n",
              "log     0.232779\n",
              "svc     0.227907\n",
              "tree    0.202703\n",
              "forest  0.151724"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 80
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ray74roTzWik",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "9b21648c-2a40-4956-b629-a84c99c62402"
      },
      "source": [
        "pd.DataFrame.from_dict(precision, orient=\"index\", columns=[\"Precision score\"])"
      ],
      "execution_count": 81,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Precision score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>perc</th>\n",
              "      <td>0.136240</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log</th>\n",
              "      <td>0.136872</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>svc</th>\n",
              "      <td>0.133515</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tree</th>\n",
              "      <td>0.176471</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>forest</th>\n",
              "      <td>0.134146</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        Precision score\n",
              "perc           0.136240\n",
              "log            0.136872\n",
              "svc            0.133515\n",
              "tree           0.176471\n",
              "forest         0.134146"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 81
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8CzVa6Dp0X6D",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "cc36d10b-00eb-4577-c581-746af256f89c"
      },
      "source": [
        "pd.DataFrame.from_dict(recall, orient=\"index\", columns=[\"Recall score\"])"
      ],
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Recall score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>perc</th>\n",
              "      <td>0.793651</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>log</th>\n",
              "      <td>0.777778</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>svc</th>\n",
              "      <td>0.777778</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tree</th>\n",
              "      <td>0.238095</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>forest</th>\n",
              "      <td>0.174603</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        Recall score\n",
              "perc        0.793651\n",
              "log         0.777778\n",
              "svc         0.777778\n",
              "tree        0.238095\n",
              "forest      0.174603"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 82
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "csqkJdMYudm4"
      },
      "source": [
        "Guardando le tabelle emerge che i modelli basati su `DecisionTreeClassifier` e `RandomForestClassifier` hanno score di determinazione e F1 score più alti rispetto agli altri modelli, tuttavia, essendo l'ambito medico il contesto di utilizzo, è più opportuno basare il nostro giudizio sul **Recall score** (ovvero i veri casi di ictus classificati correttamente).\n",
        "Ciò emerge anche dalle matrici di confusione dei vari modelli:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vRL56Rnz5Lks",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "acba4407-786e-4a9e-f38d-fe9f61dfd84e"
      },
      "source": [
        "print(\"Perceptron\")\n",
        "print(perc_conf)\n",
        "print(\"\\n\")\n",
        "print(\"Logistic Regression\")\n",
        "print(log_conf)\n",
        "print(\"\\n\")\n",
        "print(\"Support Vector Machines\")\n",
        "print(svc_conf)\n",
        "print(\"\\n\")\n",
        "print(\"Decision Tree\")\n",
        "print(tree_conf)\n",
        "print(\"\\n\")\n",
        "print(\"Random Forest\")\n",
        "print(forest_conf)"
      ],
      "execution_count": 83,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Perceptron\n",
            "           no stroke[P]  stroke[P]\n",
            "no stroke          2268        977\n",
            "stroke              572       2673\n",
            "\n",
            "\n",
            "Logistic Regression\n",
            "           no stroke[P]  stroke[P]\n",
            "no stroke          2333        912\n",
            "stroke              475       2770\n",
            "\n",
            "\n",
            "Support Vector Machines\n",
            "           no stroke[P]  stroke[P]\n",
            "no stroke          2303        942\n",
            "stroke              480       2765\n",
            "\n",
            "\n",
            "Decision Tree\n",
            "           no stroke[P]  stroke[P]\n",
            "no stroke          3049        196\n",
            "stroke              882       2363\n",
            "\n",
            "\n",
            "Random Forest\n",
            "           no stroke[P]  stroke[P]\n",
            "no stroke          3075        170\n",
            "stroke              506       2739\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Bhqd5PRw8A-2"
      },
      "source": [
        "##Confronto modelli con intervallo di confidenza\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wbV2Tcf-8GQI"
      },
      "source": [
        "Effettuiamo adesso il confronto con intervallo di confidenza al 95% per constatare quanto siano affidabili i nostri modelli migliori (`Perceptron`, `LogisticRegression` e `SVM`):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "egYQcMLhDJyj"
      },
      "source": [
        "def confidence(acc, N, Z):\n",
        "    den = (2*(N+Z**2))\n",
        "    var = (Z*np.sqrt(Z**2+4*N*acc-4*N*acc**2)) / den\n",
        "    a = (2*N*acc+Z**2) / den\n",
        "    inf = a - var\n",
        "    sup = a + var\n",
        "    return (inf, sup)\n",
        "\n",
        "def calculate_accuracy(conf_matrix):\n",
        "    return np.diag(conf_matrix).sum() / conf_matrix.sum().sum()"
      ],
      "execution_count": 84,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rO9Fi1rJHuF1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 142
        },
        "outputId": "5029fbfd-7a66-4486-b919-46229605076c"
      },
      "source": [
        "pd.DataFrame([confidence(calculate_accuracy(perc_conf), len(X_val), 1.96),\n",
        "              confidence(calculate_accuracy(log_conf), len(X_val), 1.96),\n",
        "              confidence(calculate_accuracy(svc_conf), len(X_val), 1.96)],\n",
        "                 index=[\"perceptron\", \"logistic reg\", \"SVM\"], columns=[\"inf\", \"sup\"])"
      ],
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>inf</th>\n",
              "      <th>sup</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>perceptron</th>\n",
              "      <td>0.735751</td>\n",
              "      <td>0.785147</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>logistic reg</th>\n",
              "      <td>0.761572</td>\n",
              "      <td>0.809082</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>SVM</th>\n",
              "      <td>0.755983</td>\n",
              "      <td>0.803921</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                   inf       sup\n",
              "perceptron    0.735751  0.785147\n",
              "logistic reg  0.761572  0.809082\n",
              "SVM           0.755983  0.803921"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 85
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GhpWtD-lJLE_"
      },
      "source": [
        "Come si poteva anche notare dai punteggi di recall precedentemente mostrati e dalle matrici di confusione, il modello basato su Regressione Logistica risulta il più accurato."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rYbdhXskw8zv"
      },
      "source": [
        "##Confronto con un modello casuale"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3FiAgq8TxCfQ"
      },
      "source": [
        "Addestriamo adesso un modello casuale e valutiamo come esso performa sui nostri dati:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mj9ik9Sxymgd"
      },
      "source": [
        "from sklearn.dummy import DummyClassifier"
      ],
      "execution_count": 86,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OtVs8uy9sI1x",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6a9c38e2-6ef5-43bf-8e1f-d088b72d2995"
      },
      "source": [
        "random = DummyClassifier(strategy=\"uniform\", random_state=42)\n",
        "random.fit(X_train_res, y_train_res)\n",
        "\n",
        "random_score = random.score(X_val, y_val)\n",
        "\n",
        "print(\"Score: {score}\".format(score=random_score));\n",
        "print(\"F1 score: {score}\".format(score=f1_score(y_val, random.predict(X_val))));\n",
        "print(\"Precision score: {score}\".format(score=precision_score(y_val, random.predict(X_val))))\n",
        "print(\"Recall score: {score}\".format(score=recall_score(y_val, random.predict(X_val))))\n",
        "\n",
        "scores[\"random\"] = random.score(X_val, y_val);\n",
        "f1_scores[\"random\"] = f1_score(y_val, random.predict(X_val), average=\"binary\")\n",
        "precision[\"random\"] = precision_score(y_val, random.predict(X_val))\n",
        "recall[\"random\"] = recall_score(y_val, random.predict(X_val))"
      ],
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Score: 0.5026269702276708\n",
            "F1 score: 0.10410094637223975\n",
            "Precision score: 0.05779334500875657\n",
            "Recall score: 0.5238095238095238\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8f512VF9z2eo",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "outputId": "e8d84aec-1888-4de6-f378-db6151ce0a05"
      },
      "source": [
        "random_conf = pd.DataFrame(confusion_matrix(y_res, random.predict(X_res)), columns=classes, index = index)\n",
        "random_conf"
      ],
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>no stroke[P]</th>\n",
              "      <th>stroke[P]</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>no stroke</th>\n",
              "      <td>1629</td>\n",
              "      <td>1616</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>stroke</th>\n",
              "      <td>1641</td>\n",
              "      <td>1604</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "           no stroke[P]  stroke[P]\n",
              "no stroke          1629       1616\n",
              "stroke             1641       1604"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 88
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FmcjKwh49DH4"
      },
      "source": [
        "Ovviamente i punteggi ottenuti non sono paragonabili a quelli dei modelli precedentemente addestrati.\n",
        "Procediamo confrontando il modello casuale con il modello basato su Regressione Logistica:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KPmJj_rP9JfK"
      },
      "source": [
        "from scipy import stats\n",
        "\n",
        "def difference_between_two_models(error1, error2, confidence):\n",
        "    z_half_alfa = stats.norm.ppf(confidence)\n",
        "    variance = (((1 - error1) * error1) / len(y_val)) + (((1 - error2) * error2) / len(y_val))\n",
        "    d_minus = abs(error1 - error2) - z_half_alfa * (pow(variance, 0.5))\n",
        "    d_plus = abs(error1 - error2) + z_half_alfa * (pow(variance, 0.5))\n",
        "    print(\"Valore minimo: {}\\nValore massimo: {}\\n\".format(d_minus, d_plus))"
      ],
      "execution_count": 89,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CcQCFCFS9O8h",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f67bc3ca-ef36-4a9e-f487-8562dcb1a94a"
      },
      "source": [
        "difference_between_two_models(1 - f1_scores[\"log\"], 1 - f1_scores[\"random\"], 0.99)"
      ],
      "execution_count": 90,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Valore minimo: 0.09278498157235394\n",
            "Valore massimo: 0.16457132045751333\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v0xfDXPPP_3-"
      },
      "source": [
        "##Conclusioni"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Brp2xb8TausW"
      },
      "source": [
        "Considerando i valori ottenuti dal modello, specialmente facendo riferimento allo score di recall ottenuto, mi reputo personalmente soddisfatto del lavoro svolto, poichè seppur l'F1 score ottenuto non sia molto alto rispetto a quello di altri progetti presenti su Kaggle (diversi modelli non bilanciavano correttamente la distribuzione delle classi, da qui score estremamente alti), il modello si rivela solido. Ovviamente il modello non può rimpiazzare il giudizio di un medico ma può essere un buono strumento per analizzare e valutare le condizioni di salute di un individuo rispetto alla possibilità di colpi apoplettici"
      ]
    }
  ]
}