Practical_Application_III_TB_1_FeatureEngr_Modelling_Classification_Final.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ttb-git/Berkeley-Engr-PracticalAssignment3/blob/main/Practical_Application_III_TB_1_FeatureEngr_Modelling_Classification_Final.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w_cVPvrzJ2_f"
      },
      "source": [
        "# Practical Application III: Comparing Classifiers\n",
        "\n",
        "**Overview**: In this practical application, your goal is to compare the performance of the classifiers we encountered in this section, namely K Nearest Neighbor, Logistic Regression, Decision Trees, and Support Vector Machines.  We will utilize a dataset related to marketing bank products over the telephone.  \n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gLY8yYE0J2_h"
      },
      "source": [
        "### Getting Started\n",
        "\n",
        "Our dataset comes from the UCI Machine Learning repository [link](https://archive.ics.uci.edu/ml/datasets/bank+marketing).  The data is from a Portugese banking institution and is a collection of the results of multiple marketing campaigns.  We will make use of the article accompanying the dataset [here](CRISP-DM-BANK.pdf) for more information on the data and features.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "m6vbNo22J2_l"
      },
      "source": [
        "###Understanding the Features\n",
        "\n",
        "```\n",
        "Data Set Information: The data is related with direct marketing campaigns of a Portuguese banking institution. The marketing campaigns were based on phone calls. Often, more than one contact to the same client was required, in order to access if the product (bank term deposit) would be ('yes') or not ('no') subscribed.\n",
        "\n",
        "## Attribute Information:\n",
        "\n",
        "### Bank client data:\n",
        "* Age (numeric)\n",
        "* Job : type of job (categorical: 'admin.', 'blue-collar', 'entrepreneur', 'housemaid', 'management', 'retired', 'self-employed', 'services', 'student', 'technician', 'unemployed', 'unknown')\n",
        "* Marital : marital status (categorical: 'divorced', 'married', 'single', 'unknown' ; note: 'divorced' means divorced or widowed)\n",
        "* Education (categorical: 'basic.4y', 'basic.6y', 'basic.9y', 'high.school', 'illiterate', 'professional.course', 'university.degree', 'unknown')\n",
        "* Default: has credit in default? (categorical: 'no', 'yes', 'unknown')\n",
        "* Housing: has housing loan? (categorical: 'no', 'yes', 'unknown')\n",
        "* Loan: has personal loan? (categorical: 'no', 'yes', 'unknown')\n",
        "\n",
        "###Related with the last contact of the current campaign:\n",
        "\n",
        "* Contact: contact communication type (categorical: 'cellular','telephone')\n",
        "Month: last contact month of year (categorical: 'jan', 'feb', 'mar', ..., 'nov', 'dec')\n",
        "* Day_of_week: last contact day of the week (categorical: 'mon','tue','wed','thu','fri')\n",
        "* Duration: last contact duration, in seconds (numeric). Important note: this attribute highly affects the output target (e.g., if duration=0 then y='no'). Yet, the duration is not known before a call is performed. Also, after the end of the call y is obviously known. Thus, this input should only be included for benchmark purposes and should be discarded if the intention is to have a realistic predictive model.\n",
        "\n",
        "###Other attributes:\n",
        "* Campaign: number of contacts performed during this campaign and for this client (numeric, includes last contact)\n",
        "* Pdays: number of days that passed by after the client was last contacted from a previous campaign (numeric; 999 means client was not previously contacted)\n",
        "* Previous: number of contacts performed before this campaign and for this client (numeric)\n",
        "* Poutcome: outcome of the previous marketing campaign (categorical: 'failure','nonexistent','success')\n",
        "\n",
        "### Social and economic context attributes\n",
        "* Emp.var.rate: employment variation rate - quarterly indicator (numeric)\n",
        "* Cons.price.idx: consumer price index - monthly indicator (numeric)\n",
        "* Cons.conf.idx: consumer confidence index - monthly indicator (numeric)\n",
        "* Euribor3m: euribor 3 month rate - daily indicator (numeric)\n",
        "* Nr.employed: number of employees - quarterly indicator (numeric)\n",
        "\n",
        "### Output variable (desired target):\n",
        "y - has the client subscribed a term deposit? (binary: 'yes', 'no')\n",
        "'''\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "EX3w9WKSJ2_j"
      },
      "outputs": [],
      "source": [
        "# Importing modules and libraries\n",
        "%matplotlib inline\n",
        "import os\n",
        "import pandas as pd\n",
        "import matplotlib\n",
        "matplotlib.use(u'nbAgg')\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import numpy as np\n",
        "import pickle\n",
        "from sklearn.manifold import TSNE\n",
        "from sklearn import preprocessing\n",
        "import pandas as pd\n",
        "from sklearn.linear_model import SGDClassifier\n",
        "from xgboost import XGBClassifier\n",
        "from sklearn.model_selection import RandomizedSearchCV \n",
        "\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.calibration import CalibratedClassifierCV\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "from sklearn.metrics import roc_auc_score\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.ensemble import RandomForestClassifier"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "a-uVoplZRamK",
        "outputId": "7ea878fe-72d0-4805-b9a8-a020d9684d47"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VVkJFMV5J2_k"
      },
      "outputs": [],
      "source": [
        "#Loading the dataset\n",
        "data = pd.read_csv('/content/drive/MyDrive/Berkeley-Engineering-Projects/module_17_starter/data/bank-additional-full.csv', sep = ';')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data.describe()"
      ],
      "metadata": {
        "id": "4O6YdzT13mUg",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        },
        "outputId": "269349c9-16cb-43c6-ec39-39737ecb4ff8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "               age      duration      campaign         pdays      previous  \\\n",
              "count  41176.00000  41176.000000  41176.000000  41176.000000  41176.000000   \n",
              "mean      40.02380    258.315815      2.567879    962.464810      0.173013   \n",
              "std       10.42068    259.305321      2.770318    186.937102      0.494964   \n",
              "min       17.00000      0.000000      1.000000      0.000000      0.000000   \n",
              "25%       32.00000    102.000000      1.000000    999.000000      0.000000   \n",
              "50%       38.00000    180.000000      2.000000    999.000000      0.000000   \n",
              "75%       47.00000    319.000000      3.000000    999.000000      0.000000   \n",
              "max       98.00000   4918.000000     56.000000    999.000000      7.000000   \n",
              "\n",
              "       emp.var.rate  cons.price.idx  cons.conf.idx     euribor3m   nr.employed  \n",
              "count  41176.000000    41176.000000   41176.000000  41176.000000  41176.000000  \n",
              "mean       0.081922       93.575720     -40.502863      3.621293   5167.034870  \n",
              "std        1.570883        0.578839       4.627860      1.734437     72.251364  \n",
              "min       -3.400000       92.201000     -50.800000      0.634000   4963.600000  \n",
              "25%       -1.800000       93.075000     -42.700000      1.344000   5099.100000  \n",
              "50%        1.100000       93.749000     -41.800000      4.857000   5191.000000  \n",
              "75%        1.400000       93.994000     -36.400000      4.961000   5228.100000  \n",
              "max        1.400000       94.767000     -26.900000      5.045000   5228.100000  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-1d7281e2-d541-4669-8db9-701b246fa626\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>duration</th>\n",
              "      <th>campaign</th>\n",
              "      <th>pdays</th>\n",
              "      <th>previous</th>\n",
              "      <th>emp.var.rate</th>\n",
              "      <th>cons.price.idx</th>\n",
              "      <th>cons.conf.idx</th>\n",
              "      <th>euribor3m</th>\n",
              "      <th>nr.employed</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>41176.00000</td>\n",
              "      <td>41176.000000</td>\n",
              "      <td>41176.000000</td>\n",
              "      <td>41176.000000</td>\n",
              "      <td>41176.000000</td>\n",
              "      <td>41176.000000</td>\n",
              "      <td>41176.000000</td>\n",
              "      <td>41176.000000</td>\n",
              "      <td>41176.000000</td>\n",
              "      <td>41176.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>40.02380</td>\n",
              "      <td>258.315815</td>\n",
              "      <td>2.567879</td>\n",
              "      <td>962.464810</td>\n",
              "      <td>0.173013</td>\n",
              "      <td>0.081922</td>\n",
              "      <td>93.575720</td>\n",
              "      <td>-40.502863</td>\n",
              "      <td>3.621293</td>\n",
              "      <td>5167.034870</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>10.42068</td>\n",
              "      <td>259.305321</td>\n",
              "      <td>2.770318</td>\n",
              "      <td>186.937102</td>\n",
              "      <td>0.494964</td>\n",
              "      <td>1.570883</td>\n",
              "      <td>0.578839</td>\n",
              "      <td>4.627860</td>\n",
              "      <td>1.734437</td>\n",
              "      <td>72.251364</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>17.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>-3.400000</td>\n",
              "      <td>92.201000</td>\n",
              "      <td>-50.800000</td>\n",
              "      <td>0.634000</td>\n",
              "      <td>4963.600000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>32.00000</td>\n",
              "      <td>102.000000</td>\n",
              "      <td>1.000000</td>\n",
              "      <td>999.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>-1.800000</td>\n",
              "      <td>93.075000</td>\n",
              "      <td>-42.700000</td>\n",
              "      <td>1.344000</td>\n",
              "      <td>5099.100000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>38.00000</td>\n",
              "      <td>180.000000</td>\n",
              "      <td>2.000000</td>\n",
              "      <td>999.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.100000</td>\n",
              "      <td>93.749000</td>\n",
              "      <td>-41.800000</td>\n",
              "      <td>4.857000</td>\n",
              "      <td>5191.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>47.00000</td>\n",
              "      <td>319.000000</td>\n",
              "      <td>3.000000</td>\n",
              "      <td>999.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.400000</td>\n",
              "      <td>93.994000</td>\n",
              "      <td>-36.400000</td>\n",
              "      <td>4.961000</td>\n",
              "      <td>5228.100000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>98.00000</td>\n",
              "      <td>4918.000000</td>\n",
              "      <td>56.000000</td>\n",
              "      <td>999.000000</td>\n",
              "      <td>7.000000</td>\n",
              "      <td>1.400000</td>\n",
              "      <td>94.767000</td>\n",
              "      <td>-26.900000</td>\n",
              "      <td>5.045000</td>\n",
              "      <td>5228.100000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-1d7281e2-d541-4669-8db9-701b246fa626')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-1d7281e2-d541-4669-8db9-701b246fa626 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-1d7281e2-d541-4669-8db9-701b246fa626');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%matplotlib inline\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "sns.set_theme(style=\"white\")\n",
        "\n",
        "# Computing the correlation matrix\n",
        "corr = data.corr()\n",
        "\n",
        "# Generating a mask for the upper triangle\n",
        "mask = np.triu(np.ones_like(corr, dtype=bool))\n",
        "\n",
        "# Setting up the matplotlib figure\n",
        "f, ax = plt.subplots(figsize=(11, 9))\n",
        "\n",
        "# Generating a custom diverging colormap\n",
        "cmap = sns.diverging_palette(230, 20, as_cmap=True)\n",
        "\n",
        "# Drawing the heatmap with the mask and correct aspect ratio\n",
        "sns.heatmap(corr, mask=mask, cmap=cmap, vmax=.3, center=0, square=True, linewidths=.5, cbar_kws={\"shrink\": .5})"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 623
        },
        "id": "KuQmOupzU8no",
        "outputId": "b8361ade-251c-430d-913b-ec5bc995d0e3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f576435de50>"
            ]
          },
          "metadata": {},
          "execution_count": 196
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 792x648 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9sDgZ6vvJ2_o"
      },
      "source": [
        "### Problem 5: Engineering Features\n",
        "\n",
        "Now that you understand your business objective, we will build a basic model to get started.  Before we can do this, we must work to encode the data.  Using just the bank information features (columns 1 - 7), prepare the features and target column for modeling with appropriate encoding and transformations."
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Let's drop the duplicate values first\n",
        "data = data.drop_duplicates()\n",
        "data.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SGpbcxFMe12a",
        "outputId": "a03898fa-2196-4e88-c24d-77e07ac19c78"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(41176, 21)"
            ]
          },
          "metadata": {},
          "execution_count": 197
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "####Dealing with Missing Data and Duplicate Values\n"
      ],
      "metadata": {
        "id": "GhM7wR0ACofz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Checking for Null Values\n",
        "data.isnull().any()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dBSNeaqrLA0Z",
        "outputId": "01346f18-c41c-4373-b605-5b9df0276732"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "age               False\n",
              "job               False\n",
              "marital           False\n",
              "education         False\n",
              "default           False\n",
              "housing           False\n",
              "loan              False\n",
              "contact           False\n",
              "month             False\n",
              "day_of_week       False\n",
              "duration          False\n",
              "campaign          False\n",
              "pdays             False\n",
              "previous          False\n",
              "poutcome          False\n",
              "emp.var.rate      False\n",
              "cons.price.idx    False\n",
              "cons.conf.idx     False\n",
              "euribor3m         False\n",
              "nr.employed       False\n",
              "y                 False\n",
              "dtype: bool"
            ]
          },
          "metadata": {},
          "execution_count": 198
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We have no null values!"
      ],
      "metadata": {
        "id": "73wk5ED5ntyl"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "###I am going to select bank information features (columns 1 - 7) and prepare the features and target column for modeling with appropriate encoding and transformations.\n"
      ],
      "metadata": {
        "id": "YlofRaAyH8gU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_feature = data.iloc[: , 0:7]\n",
        "df_feature.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "6DySJQnGH9C3",
        "outputId": "08d1fda1-aa9c-4c0d-dcf8-f20e785d245a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   age        job  marital    education  default housing loan\n",
              "0   56  housemaid  married     basic.4y       no      no   no\n",
              "1   57   services  married  high.school  unknown      no   no\n",
              "2   37   services  married  high.school       no     yes   no\n",
              "3   40     admin.  married     basic.6y       no      no   no\n",
              "4   56   services  married  high.school       no      no  yes"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-bbf1da6a-3297-4843-a5a6-7d155e69964c\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>job</th>\n",
              "      <th>marital</th>\n",
              "      <th>education</th>\n",
              "      <th>default</th>\n",
              "      <th>housing</th>\n",
              "      <th>loan</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>56</td>\n",
              "      <td>housemaid</td>\n",
              "      <td>married</td>\n",
              "      <td>basic.4y</td>\n",
              "      <td>no</td>\n",
              "      <td>no</td>\n",
              "      <td>no</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>57</td>\n",
              "      <td>services</td>\n",
              "      <td>married</td>\n",
              "      <td>high.school</td>\n",
              "      <td>unknown</td>\n",
              "      <td>no</td>\n",
              "      <td>no</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>37</td>\n",
              "      <td>services</td>\n",
              "      <td>married</td>\n",
              "      <td>high.school</td>\n",
              "      <td>no</td>\n",
              "      <td>yes</td>\n",
              "      <td>no</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>40</td>\n",
              "      <td>admin.</td>\n",
              "      <td>married</td>\n",
              "      <td>basic.6y</td>\n",
              "      <td>no</td>\n",
              "      <td>no</td>\n",
              "      <td>no</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>56</td>\n",
              "      <td>services</td>\n",
              "      <td>married</td>\n",
              "      <td>high.school</td>\n",
              "      <td>no</td>\n",
              "      <td>no</td>\n",
              "      <td>yes</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-bbf1da6a-3297-4843-a5a6-7d155e69964c')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-bbf1da6a-3297-4843-a5a6-7d155e69964c button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-bbf1da6a-3297-4843-a5a6-7d155e69964c');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 199
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Encoding Categorical features\n",
        "Jobs, Marital, Education, Default, Housing, Loan. Converting to continuous due the feature scaling that will be applied later."
      ],
      "metadata": {
        "id": "A6h80tYqsISe"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Label encoder order is alphabetical\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "labelencoder_X = LabelEncoder()\n",
        "df_feature['job']      = labelencoder_X.fit_transform(df_feature['job']) \n",
        "df_feature['marital']  = labelencoder_X.fit_transform(df_feature['marital']) \n",
        "df_feature['education']= labelencoder_X.fit_transform(df_feature['education']) \n",
        "df_feature['default']  = labelencoder_X.fit_transform(df_feature['default']) \n",
        "df_feature['housing']  = labelencoder_X.fit_transform(df_feature['housing']) \n",
        "df_feature['loan']     = labelencoder_X.fit_transform(df_feature['loan']) "
      ],
      "metadata": {
        "id": "hgqsv9LVsI0T"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#function to creat group of ages, this helps because we have 78 differente values here\n",
        "def age(dataframe):\n",
        "    dataframe.loc[dataframe['age'] <= 32, 'age'] = 1\n",
        "    dataframe.loc[(dataframe['age'] > 32) & (dataframe['age'] <= 47), 'age'] = 2\n",
        "    dataframe.loc[(dataframe['age'] > 47) & (dataframe['age'] <= 70), 'age'] = 3\n",
        "    dataframe.loc[(dataframe['age'] > 70) & (dataframe['age'] <= 98), 'age'] = 4\n",
        "           \n",
        "    return dataframe\n",
        "\n",
        "age(df_feature);"
      ],
      "metadata": {
        "id": "jgLi2iM_qKHi"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(df_feature.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0QczwnJEsJDt",
        "outputId": "ec818fe4-a3fa-42a8-9d4e-8cd57055226e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(41176, 7)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X = df_feature\n",
        "X.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "IDRb7bvizhbO",
        "outputId": "05319dc0-4dda-4f91-f35e-5b39c757cd1b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   age  job  marital  education  default  housing  loan\n",
              "0    3    3        1          0        0        0     0\n",
              "1    3    7        1          3        1        0     0\n",
              "2    2    7        1          3        0        2     0\n",
              "3    2    0        1          1        0        0     0\n",
              "4    3    7        1          3        0        0     2"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-17dced8d-3331-4d43-85d1-b44eb331d892\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>job</th>\n",
              "      <th>marital</th>\n",
              "      <th>education</th>\n",
              "      <th>default</th>\n",
              "      <th>housing</th>\n",
              "      <th>loan</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>3</td>\n",
              "      <td>7</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>7</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>3</td>\n",
              "      <td>7</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-17dced8d-3331-4d43-85d1-b44eb331d892')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-17dced8d-3331-4d43-85d1-b44eb331d892 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-17dced8d-3331-4d43-85d1-b44eb331d892');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 203
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data.apply(lambda x: len(x.unique()), axis=0)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BzSnb0eq44EC",
        "outputId": "82b4428d-64f6-461d-f2a5-f625232d39ce"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "age                 78\n",
              "job                 12\n",
              "marital              4\n",
              "education            8\n",
              "default              3\n",
              "housing              3\n",
              "loan                 3\n",
              "contact              2\n",
              "month               10\n",
              "day_of_week          5\n",
              "duration          1544\n",
              "campaign            42\n",
              "pdays               27\n",
              "previous             8\n",
              "poutcome             3\n",
              "emp.var.rate        10\n",
              "cons.price.idx      26\n",
              "cons.conf.idx       26\n",
              "euribor3m          316\n",
              "nr.employed         11\n",
              "y                    2\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 204
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Normalizig the Data\n",
        "Data Standardization give data zero mean and unit variance (technically should be done after train test split )"
      ],
      "metadata": {
        "id": "SV2nyuci0TY6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Converting dependent variable categorical to dummy\n",
        "y = pd.get_dummies(data['y'], columns = ['y'], prefix = ['y'], drop_first = True)"
      ],
      "metadata": {
        "id": "78rT1Uxyt3oT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y.info()\n",
        "print(y.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "x6UK-cGet3xZ",
        "outputId": "886c0537-0704-47fc-fc13-ad468e466851"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 41176 entries, 0 to 41187\n",
            "Data columns (total 1 columns):\n",
            " #   Column     Non-Null Count  Dtype\n",
            "---  ------     --------------  -----\n",
            " 0   ['y']_yes  41176 non-null  uint8\n",
            "dtypes: uint8(1)\n",
            "memory usage: 361.9 KB\n",
            "(41176, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X= preprocessing.StandardScaler().fit(X).transform(X)\n",
        "X[0:5]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gFE_ylNH06H_",
        "outputId": "63a2d628-f21f-4280-c50a-9b555c3c535e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 1.38828682, -0.20147477, -0.28370335, -1.7537527 , -0.51365651,\n",
              "        -1.087787  , -0.45257005],\n",
              "       [ 1.38828682,  0.9113627 , -0.28370335, -0.34959564,  1.94511127,\n",
              "        -1.087787  , -0.45257005],\n",
              "       [ 0.0290815 ,  0.9113627 , -0.28370335, -0.34959564, -0.51365651,\n",
              "         0.94206515, -0.45257005],\n",
              "       [ 0.0290815 , -1.03610287, -0.28370335, -1.28570035, -0.51365651,\n",
              "        -1.087787  , -0.45257005],\n",
              "       [ 1.38828682,  0.9113627 , -0.28370335, -0.34959564, -0.51365651,\n",
              "        -1.087787  ,  2.31104027]])"
            ]
          },
          "metadata": {},
          "execution_count": 252
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "XWprpRCOtdWI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        ""
      ],
      "metadata": {
        "id": "51nBxg_l3DVh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def evaluate_model(model, x_test, y_test):\n",
        "    from sklearn import metrics\n",
        "\n",
        "    # Predict Test Data \n",
        "    y_pred = model.predict(x_test)\n",
        "\n",
        "    # Calculate accuracy, precision, recall, f1-score, and kappa score\n",
        "    acc = metrics.accuracy_score(y_test, y_pred)\n",
        "    prec = metrics.precision_score(y_test, y_pred)\n",
        "    rec = metrics.recall_score(y_test, y_pred)\n",
        "    f1 = metrics.f1_score(y_test, y_pred)\n",
        "    kappa = metrics.cohen_kappa_score(y_test, y_pred)\n",
        "\n",
        "    # Calculate area under curve (AUC)\n",
        "    y_pred_proba = model.predict_proba(x_test)[::,1]\n",
        "    fpr, tpr, _ = metrics.roc_curve(y_test, y_pred_proba)\n",
        "    auc = metrics.roc_auc_score(y_test, y_pred_proba)\n",
        "\n",
        "    # Display confussion matrix\n",
        "    cm = metrics.confusion_matrix(y_test, y_pred)\n",
        "\n",
        "    return {'acc': acc, 'prec': prec, 'rec': rec, 'f1': f1, 'kappa': kappa, \n",
        "            'fpr': fpr, 'tpr': tpr, 'auc': auc, 'cm': cm}"
      ],
      "metadata": {
        "id": "cjppQImF3Du9"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8562km4MJ2_p"
      },
      "source": [
        "### Problem 6: Train/Test Split\n",
        "\n",
        "With your data prepared, split it into a train and test set."
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Classification"
      ],
      "metadata": {
        "id": "pv27Z6fkAXvG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Separate into train and test sets\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "# Separate train and test sets in a balanced way\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.15, random_state=42, stratify= y)\n",
        "print ('Train set:', X_train.shape, y_train.shape)\n",
        "print ('Test set:', X_test.shape,  y_test.shape)\n",
        "\n",
        "#print(f'\\n\\n{features_train[0]}\\n\\n')\n",
        "\n",
        "unique, counts = np.unique(y, return_counts=True)\n",
        "print(f'Total set distribution: {dict(zip(unique, counts))}')\n",
        "\n",
        "unique, counts = np.unique(y_train, return_counts=True)\n",
        "print(f'Train set distribution: {dict(zip(unique, counts))}')\n",
        "\n",
        "unique, counts = np.unique(y_test, return_counts=True)\n",
        "print(f'Test set distribution: {dict(zip(unique, counts))}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2kHY7llu06WZ",
        "outputId": "6417a819-eb1e-4620-b400-ef2c33d3feb6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train set: (34999, 7) (34999, 1)\n",
            "Test set: (6177, 7) (6177, 1)\n",
            "Total set distribution: {0: 36537, 1: 4639}\n",
            "Train set distribution: {0: 31056, 1: 3943}\n",
            "Test set distribution: {0: 5481, 1: 696}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "OZUaXtuYWQo3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Problem 7: A Baseline Model\n",
        "\n",
        "Before we build our first model, we want to establish a baseline.  What is the baseline performance that our classifier should aim to beat?"
      ],
      "metadata": {
        "id": "Tfab-LcuIMPG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import time\n",
        "from sklearn.metrics import roc_auc_score\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "# Initial simple model with logistic regression\n",
        "start_time = time.time()##\n",
        "\n",
        "lgr = LogisticRegression(C=0.001)\n",
        "lgr.fit(X_train, y_train)\n",
        "\n",
        "y_pred = lgr.predict(X_test)\n",
        "lgr.score(X_test, y_test)\n",
        "\n",
        "# Evaluate Model\n",
        "lgr_eval = evaluate_model(lgr, X_test, y_test)\n",
        "\n",
        "end_time = time.time()##"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MDgKRN59IQoG",
        "outputId": "ff264d37-cdcb-4ff3-b6d4-3354022ecb01"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:993: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we see above, the AUC score is 0.62."
      ],
      "metadata": {
        "id": "H6WVSeQdKaNm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Print result\n",
        "\n",
        "print(\"\\nRun time for baseline model : \", end_time-start_time)\n",
        "print('Accuracy:', lgr_eval['acc'])\n",
        "print('Precision:', lgr_eval['prec'])\n",
        "print('Recall:', lgr_eval['rec'])\n",
        "print('F1 Score:', lgr_eval['f1'])\n",
        "print('Cohens Kappa Score:', lgr_eval['kappa'])\n",
        "print('Area Under Curve:', lgr_eval['auc'])\n",
        "print('Confusion Matrix:\\n', lgr_eval['cm'])\n",
        "\n",
        "print(metrics.classification_report(y_test, y_pred))"
      ],
      "metadata": {
        "id": "xAF_iBi9_Wsn",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fd566815-2dde-4e88-c1ab-ebe38b55afc5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Run time for baseline model :  0.09002113342285156\n",
            "Accuracy: 0.8873239436619719\n",
            "Precision: 0.0\n",
            "Recall: 0.0\n",
            "F1 Score: 0.0\n",
            "Cohens Kappa Score: 0.0\n",
            "Area Under Curve: 0.6216454124698279\n",
            "Confusion Matrix:\n",
            " [[5481    0]\n",
            " [ 696    0]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.89      1.00      0.94      5481\n",
            "           1       0.00      0.00      0.00       696\n",
            "\n",
            "    accuracy                           0.89      6177\n",
            "   macro avg       0.44      0.50      0.47      6177\n",
            "weighted avg       0.79      0.89      0.83      6177\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Problem 8: A Simple Model\n",
        "Use Logistic Regression to build a basic model on your data."
      ],
      "metadata": {
        "id": "0hlCMn6NLRbk"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### A Simple model\n",
        "\n",
        "I will create a simple LogisticRegression model without any hyper-parameter tuning and apply that to the data first."
      ],
      "metadata": {
        "id": "QEFY71WZJrr7"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Initial simple model with logistic regression\n",
        "start_time = time.time()##\n",
        "\n",
        "lr = LogisticRegression(C=0.001)\n",
        "lr.fit(X_train, y_train)\n",
        "\n",
        "y_pred = lr.predict(X_test)\n",
        "lr.score(X_test, y_test)\n",
        "\n",
        "end_time = time.time()##\n",
        "\n",
        "# Evaluate Model\n",
        "lr_eval = evaluate_model(lr, X_test, y_test)\n",
        "\n",
        "# Print result\n",
        "\n",
        "print(\"\\nRun time for baseline model : \", end_time-start_time)\n",
        "print('Accuracy:', lr_eval['acc'])\n",
        "print('Precision:', lr_eval['prec'])\n",
        "print('Recall:', lr_eval['rec'])\n",
        "print('F1 Score:', lr_eval['f1'])\n",
        "print('Cohens Kappa Score:', lr_eval['kappa'])\n",
        "print('Area Under Curve:', lr_eval['auc'])\n",
        "print('Confusion Matrix:\\n', lr_eval['cm'])\n",
        "\n",
        "print(metrics.classification_report(y_test, y_pred))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JS0UEUVo_V7C",
        "outputId": "89b06c8b-8a57-4283-cf4d-e6ea7f3cdec3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Run time for baseline model :  0.07568192481994629\n",
            "Accuracy: 0.8873239436619719\n",
            "Precision: 0.0\n",
            "Recall: 0.0\n",
            "F1 Score: 0.0\n",
            "Cohens Kappa Score: 0.0\n",
            "Area Under Curve: 0.6216454124698279\n",
            "Confusion Matrix:\n",
            " [[5481    0]\n",
            " [ 696    0]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.89      1.00      0.94      5481\n",
            "           1       0.00      0.00      0.00       696\n",
            "\n",
            "    accuracy                           0.89      6177\n",
            "   macro avg       0.44      0.50      0.47      6177\n",
            "weighted avg       0.79      0.89      0.83      6177\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:993: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "3qhEW2mE_WFB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Problem 9: Score the Model\n",
        "\n",
        "What is the accuracy of your model?"
      ],
      "metadata": {
        "id": "bMdgdARpLo7g"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import sklearn.metrics as metrics\n",
        "from sklearn.metrics import roc_auc_score\n",
        "from sklearn.metrics import roc_curve\n",
        "\n",
        "start_time = time.time()##\n",
        "\n",
        "metrics.confusion_matrix(y_test, y_pred)\n",
        "print(metrics.classification_report(y_test, y_pred))\n",
        "\n",
        "end_time = time.time()##\n",
        "\n",
        "# Print result\n",
        "\n",
        "print(\"\\nRun time for baseline model : \", end_time-start_time)\n",
        "print('Accuracy:', dtc_eval['acc'])\n",
        "print('Precision:', dtc_eval['prec'])\n",
        "print('Recall:', dtc_eval['rec'])\n",
        "print('F1 Score:', dtc_eval['f1'])\n",
        "print('Cohens Kappa Score:', dtc_eval['kappa'])\n",
        "print('Area Under Curve:', dtc_eval['auc'])\n",
        "print('Confusion Matrix:\\n', dtc_eval['cm'])\n",
        "\n",
        "print(metrics.classification_report(y_test, y_pred))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fc3We6tFanNV",
        "outputId": "d0035d16-1c53-49dc-c34b-55a67de1ca14"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.89      1.00      0.94      5481\n",
            "           1       0.00      0.00      0.00       696\n",
            "\n",
            "    accuracy                           0.89      6177\n",
            "   macro avg       0.44      0.50      0.47      6177\n",
            "weighted avg       0.79      0.89      0.83      6177\n",
            "\n",
            "\n",
            "Run time for baseline model :  0.08014893531799316\n",
            "Accuracy: 0.8831147806378501\n",
            "Precision: 0.2903225806451613\n",
            "Recall: 0.02586206896551724\n",
            "F1 Score: 0.0474934036939314\n",
            "Cohens Kappa Score: 0.029606633097338575\n",
            "Area Under Curve: 0.6213262587370793\n",
            "Confusion Matrix:\n",
            " [[5437   44]\n",
            " [ 678   18]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.89      1.00      0.94      5481\n",
            "           1       0.00      0.00      0.00       696\n",
            "\n",
            "    accuracy                           0.89      6177\n",
            "   macro avg       0.44      0.50      0.47      6177\n",
            "weighted avg       0.79      0.89      0.83      6177\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plotting\n",
        "probs = model_base.predict_proba(X_test)\n",
        "preds = probs[:,1]\n",
        "fpr, tpr, threshold = metrics.roc_curve(y_test, preds)\n",
        "logit_roc_auc = metrics.auc(fpr, tpr)\n",
        "\n",
        "plt.plot(fpr, tpr, label='Logistic Regression Base (area = %0.2f)' % logit_roc_auc)\n",
        "plt.plot([0, 1], [0, 1],'r--')\n",
        "plt.xlim([0.0, 1.0])\n",
        "plt.ylim([0.0, 1.05])\n",
        "plt.xlabel('False Positive Rate')\n",
        "plt.ylabel('True Positive Rate')\n",
        "plt.title('Receiver operating characteristic')\n",
        "plt.legend(loc=\"lower right\")\n",
        "plt.savefig('Log_ROC')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "id": "dZKP4Hr94NLl",
        "outputId": "5b8589e5-e426-4a79-811f-28cdfe981f8b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "rgHEv_507vxq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Problem 10: Model Comparisons\n",
        "\n",
        "Now, we aim to compare the performance of the Logistic Regression model to our KNN algorithm, Decision Tree, and SVM models.  Using the default settings for each of the models, fit and score each.  Also, be sure to compare the fit time of each of the models.  Present your findings in a `DataFrame` similar to that below:\n",
        "\n",
        "| Model | Train Time | Train Accuracy | Test Accuracy |\n",
        "| ----- | ---------- | -------------  | -----------   |\n",
        "|     |    |.     |.     |"
      ],
      "metadata": {
        "id": "bj_09Hj67xR4"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## KNN"
      ],
      "metadata": {
        "id": "F9R6G9ef0x9N"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Import knearest neighbors Classifier model\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "\n",
        "#Create KNN Classifier\n",
        "knn = KNeighborsClassifier(n_neighbors=5)\n",
        "\n",
        "start_time = time.time()##\n",
        "\n",
        "#Train the model using the training sets\n",
        "knn.fit(X_train, y_train)\n",
        "\n",
        "#Predict the response for test dataset\n",
        "y_pred = knn.predict(X_test)\n",
        "\n",
        "# Evaluate Model\n",
        "knn_eval = evaluate_model(nb, X_test, y_test)\n",
        "\n",
        "end_time = time.time()##\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xN0Za2nP0vJu",
        "outputId": "bdc4b590-f0a9-4a12-c00f-d89af5bd4660"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/neighbors/_classification.py:198: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
            "  return self._fit(X, y)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Print result\n",
        "\n",
        "print(\"\\nRun time for baseline model : \", end_time-start_time)\n",
        "print('Accuracy:', knn_eval['acc'])\n",
        "print('Precision:', knn_eval['prec'])\n",
        "print('Recall:', knn_eval['rec'])\n",
        "print('F1 Score:', knn_eval['f1'])\n",
        "print('Cohens Kappa Score:', knn_eval['kappa'])\n",
        "print('Area Under Curve:', knn_eval['auc'])\n",
        "print('Confusion Matrix:\\n', knn_eval['cm'])\n",
        "\n",
        "print(metrics.classification_report(y_test, y_pred))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XJX6ESnT1LvI",
        "outputId": "90060b4c-348b-4b72-9aba-2c2556b017e0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Run time for baseline model :  0.9117605686187744\n",
            "Accuracy: 0.8834272391684477\n",
            "Precision: 0.0\n",
            "Recall: 0.0\n",
            "F1 Score: 0.0\n",
            "Cohens Kappa Score: 0.0\n",
            "Area Under Curve: 0.6434817370427388\n",
            "Confusion Matrix:\n",
            " [[9094    0]\n",
            " [1200    0]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.89      0.99      0.93      9094\n",
            "           1       0.33      0.05      0.09      1200\n",
            "\n",
            "    accuracy                           0.88     10294\n",
            "   macro avg       0.61      0.52      0.51     10294\n",
            "weighted avg       0.82      0.88      0.84     10294\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We got a classification rate of 88%,  It is considered as good accuracy.\n",
        "\n",
        "For further evaluation, you we will create a model for a different number of neighbors later."
      ],
      "metadata": {
        "id": "bK1bcRIW2JDn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "YXjCl4zp1L2e"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Decision Tree"
      ],
      "metadata": {
        "id": "CqsHzrnOkUem"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn import tree\n",
        "\n",
        "start_time = time.time()##\n",
        "# Building Decision Tree model \n",
        "dtc = tree.DecisionTreeClassifier(random_state=0)\n",
        "dtc.fit(X_train, y_train)\n"
      ],
      "metadata": {
        "id": "IB5F9-fx7v9H",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "74e72aff-c78a-4466-a01e-c4157ed9d5b3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "DecisionTreeClassifier(random_state=0)"
            ]
          },
          "metadata": {},
          "execution_count": 290
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate Model\n",
        "dtc_eval = evaluate_model(dtc, X_test, y_test)\n",
        "\n",
        "end_time = time.time()##\n",
        "\n",
        "# Print result\n",
        "\n",
        "print(\"\\nRun time for baseline model : \", end_time-start_time)\n",
        "print('Accuracy:', dtc_eval['acc'])\n",
        "print('Precision:', dtc_eval['prec'])\n",
        "print('Recall:', dtc_eval['rec'])\n",
        "print('F1 Score:', dtc_eval['f1'])\n",
        "print('Cohens Kappa Score:', dtc_eval['kappa'])\n",
        "print('Area Under Curve:', dtc_eval['auc'])\n",
        "print('Confusion Matrix:\\n', dtc_eval['cm'])\n",
        "\n",
        "print(metrics.classification_report(y_test, y_pred))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UphbZfjJFf_Q",
        "outputId": "1fcee255-5063-4918-e8a5-362144bb07d1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Run time for baseline model :  3.6465373039245605\n",
            "Accuracy: 0.87759860112687\n",
            "Precision: 0.20588235294117646\n",
            "Recall: 0.0175\n",
            "F1 Score: 0.03225806451612903\n",
            "Cohens Kappa Score: 0.01425354697085901\n",
            "Area Under Curve: 0.6248195696796424\n",
            "Confusion Matrix:\n",
            " [[9013   81]\n",
            " [1179   21]]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.89      0.99      0.93      9094\n",
            "           1       0.33      0.05      0.09      1200\n",
            "\n",
            "    accuracy                           0.88     10294\n",
            "   macro avg       0.61      0.52      0.51     10294\n",
            "weighted avg       0.82      0.88      0.84     10294\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "aw_WQgeBlH9q"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Support Vector Machines"
      ],
      "metadata": {
        "id": "PTELYJ54MvC9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.svm import SVC\n",
        "svc = SVC(class_weight='balanced').fit(X_train, y_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "THDQOX7VlIGf",
        "outputId": "f0668d6c-7719-423e-f353-ff27ad96c19d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:993: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "svc_train_preds = svc.predict(X_train)\n",
        "svc_test_preds = svc.predict(X_test)\n",
        "\n",
        "start_time = time.time()##\n",
        "print('Classification report for balanced classes \\n',classification_report(y_test,svc_test_preds))\n",
        "svc_test_preds = np.where(svc_test_preds == True, 1, 0)\n",
        "svc_train_preds = np.where(svc_train_preds == True, 1, 0)\n",
        "\n",
        "svc_test_mse = mean_squared_error(svc_test_preds, y_test)\n",
        "svc_train_mse = mean_squared_error(svc_train_preds, y_train)\n",
        "\n",
        "print(\"KNN Train MSE:\", svc_train_mse)\n",
        "print(\"KNN Test MSE:\", svc_test_mse)\n",
        "\n",
        "end_time = time.time()##\n",
        "\n",
        "print(\"\\nRun time for baseline model : \", end_time-start_time)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OnIJ9QddlIL3",
        "outputId": "dcf0d11f-bc3e-43ac-e9ff-d51ccc858717"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification report for balanced classes \n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.93      0.53      0.67      9094\n",
            "           1       0.16      0.69      0.26      1200\n",
            "\n",
            "    accuracy                           0.54     10294\n",
            "   macro avg       0.54      0.61      0.47     10294\n",
            "weighted avg       0.84      0.54      0.62     10294\n",
            "\n",
            "KNN Train MSE: 0.4599766854478337\n",
            "KNN Test MSE: 0.4553137750145716\n",
            "\n",
            "Run time for baseline model :  0.027289867401123047\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "5LIHxlqJlIZV"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Random Forest"
      ],
      "metadata": {
        "id": "ODgoujmkPIHq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.ensemble import RandomForestClassifier\n",
        "\n",
        "# Building Random Forest model\n",
        "start_time = time.time()##\n",
        "rf = RandomForestClassifier(random_state=0)\n",
        "rf.fit(X_train, y_train)\n",
        "\n",
        "# Evaluate Model\n",
        "rf_eval = evaluate_model(rf, X_test, y_test)\n",
        "\n",
        "end_time = time.time()##"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eUHhnHKrPHJG",
        "outputId": "129a5760-91bf-42f4-afbc-898ea3fc0b1a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:6: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
            "  \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Print result\n",
        "\n",
        "print(\"\\nRun time for baseline model : \", end_time-start_time)\n",
        "print('Accuracy:', rf_eval['acc'])\n",
        "print('Precision:', rf_eval['prec'])\n",
        "print('Recall:', rf_eval['rec'])\n",
        "print('F1 Score:', rf_eval['f1'])\n",
        "print('Cohens Kappa Score:', rf_eval['kappa'])\n",
        "print('Area Under Curve:', rf_eval['auc'])\n",
        "print('Confusion Matrix:\\n', rf_eval['cm'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3cgkh3RSQ2vw",
        "outputId": "908d9120-e754-4712-b49c-46fef9f28bdc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Run time for baseline model :  1.996901035308838\n",
            "Accuracy: 0.8785700408004663\n",
            "Precision: 0.2807017543859649\n",
            "Recall: 0.02666666666666667\n",
            "F1 Score: 0.04870624048706241\n",
            "Cohens Kappa Score: 0.029066947484575834\n",
            "Area Under Curve: 0.6275653361190529\n",
            "Confusion Matrix:\n",
            " [[9012   82]\n",
            " [1168   32]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "Digt4vAqUPXZ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Naive Bayes\n"
      ],
      "metadata": {
        "id": "zxMNCT-kUPnd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.naive_bayes import GaussianNB\n",
        "\n",
        "# Building Naive Bayes model \n",
        "start_time = time.time()##\n",
        "nb = GaussianNB()\n",
        "nb.fit(X_train, y_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Z3TlfIq8UP1w",
        "outputId": "aa319bcd-2ab5-4a14-8a17-a9a2509efe0f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:993: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GaussianNB()"
            ]
          },
          "metadata": {},
          "execution_count": 309
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate Model\n",
        "nb_eval = evaluate_model(nb, X_test, y_test)\n",
        "\n",
        "end_time = time.time()##\n",
        "\n",
        "# Print result\n",
        "print('Accuracy:', nb_eval['acc'])\n",
        "print('Precision:', nb_eval['prec'])\n",
        "print('Recall:', nb_eval['rec'])\n",
        "print('F1 Score:', nb_eval['f1'])\n",
        "print('Cohens Kappa Score:', nb_eval['kappa'])\n",
        "print('Area Under Curve:', nb_eval['auc'])\n",
        "print('Confusion Matrix:\\n', nb_eval['cm'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OtJqqWugUP6t",
        "outputId": "ce4b7141-5756-43ea-aba1-8d4668bc266a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy: 0.8834272391684477\n",
            "Precision: 0.0\n",
            "Recall: 0.0\n",
            "F1 Score: 0.0\n",
            "Cohens Kappa Score: 0.0\n",
            "Area Under Curve: 0.641147414045891\n",
            "Confusion Matrix:\n",
            " [[9094    0]\n",
            " [1200    0]]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model Comparison"
      ],
      "metadata": {
        "id": "HW70woQNP449"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Intitialize figure with two plots\n",
        "fig, (ax1, ax2) = plt.subplots(1, 2)\n",
        "fig.suptitle('Model Comparison', fontsize=16, fontweight='bold')\n",
        "fig.set_figheight(7)\n",
        "fig.set_figwidth(14)\n",
        "fig.set_facecolor('white')\n",
        "\n",
        "# First plot\n",
        "## set bar size\n",
        "barWidth = 0.2\n",
        "\n",
        "lgr_score = [lgr_eval['acc'], lgr_eval['prec'], lgr_eval['rec'], lgr_eval['f1'], lgr_eval['kappa']]\n",
        "dtc_score = [dtc_eval['acc'], dtc_eval['prec'], dtc_eval['rec'], dtc_eval['f1'], dtc_eval['kappa']]\n",
        "rf_score = [rf_eval['acc'], rf_eval['prec'], rf_eval['rec'], rf_eval['f1'], rf_eval['kappa']]\n",
        "nb_score = [nb_eval['acc'], nb_eval['prec'], nb_eval['rec'], nb_eval['f1'], nb_eval['kappa']]\n",
        "knn_score = [knn_eval['acc'], knn_eval['prec'], knn_eval['rec'], knn_eval['f1'], knn_eval['kappa']]\n",
        "svc_score = [svc_eval['acc'], svc_eval['prec'], svc_eval['rec'], svc_eval['f1'], svc_eval['kappa']]\n",
        "\n",
        "\n",
        "## Set position of bar on X axis\n",
        "r1 = np.arange(len(dtc_score))\n",
        "r2 = [x + barWidth for x in r1]\n",
        "r3 = [x + barWidth for x in r2]\n",
        "r4 = [x + barWidth for x in r3]\n",
        "\n",
        "## Make the plot\n",
        "ax1.bar(r3, lgr_score, width=barWidth, edgecolor='white', label='Logistic Regression')\n",
        "ax1.bar(r4, svc_score, width=barWidth, edgecolor='white', label='Support Vector Machine (SVM)')\n",
        "ax1.bar(r1, dtc_score, width=barWidth, edgecolor='white', label='Decision Tree')\n",
        "ax1.bar(r2, rf_score, width=barWidth, edgecolor='white', label='Random Forest')\n",
        "ax1.bar(r3, nb_score, width=barWidth, edgecolor='white', label='Naive Bayes')\n",
        "ax1.bar(r4, knn_score, width=barWidth, edgecolor='white', label='K-Nearest Neighbors')\n",
        "\n",
        "## Configure x and y axis\n",
        "ax1.set_xlabel('Metrics', fontweight='bold')\n",
        "labels = ['Accuracy', 'Precision', 'Recall', 'F1', 'Kappa']\n",
        "ax1.set_xticks([r + (barWidth * 1.5) for r in range(len(dtc_score))], )\n",
        "ax1.set_xticklabels(labels)\n",
        "ax1.set_ylabel('Score', fontweight='bold')\n",
        "ax1.set_ylim(0, 1)\n",
        "\n",
        "## Create legend & title\n",
        "ax1.set_title('Evaluation Metrics', fontsize=14, fontweight='bold')\n",
        "ax1.legend()\n",
        "\n",
        "# Second plot\n",
        "## Comparing ROC Curve\n",
        "\n",
        "ax2.plot(lgr_eval['fpr'], lgr_eval['tpr'], label='Logistic Regression, auc = {:0.5f}'.format(lgr_eval['auc']))\n",
        "ax2.plot(svc_eval['fpr'], svc_eval['tpr'], label='Support Vector Machines, auc = {:0.5f}'.format(rf_eval['auc']))\n",
        "ax2.plot(dtc_eval['fpr'], dtc_eval['tpr'], label='Decision Tree, auc = {:0.5f}'.format(dtc_eval['auc']))\n",
        "ax2.plot(rf_eval['fpr'], rf_eval['tpr'], label='Random Forest, auc = {:0.5f}'.format(rf_eval['auc']))\n",
        "ax2.plot(nb_eval['fpr'], nb_eval['tpr'], label='Naive Bayes, auc = {:0.5f}'.format(nb_eval['auc']))\n",
        "ax2.plot(knn_eval['fpr'], knn_eval['tpr'], label='K-Nearest Nieghbor, auc = {:0.5f}'.format(knn_eval['auc']))\n",
        "\n",
        "## Configure x and y axis\n",
        "ax2.set_xlabel('False Positive Rate', fontweight='bold')\n",
        "ax2.set_ylabel('True Positive Rate', fontweight='bold')\n",
        "\n",
        "## Create legend & title\n",
        "ax2.set_title('ROC Curve', fontsize=14, fontweight='bold')\n",
        "ax2.legend(loc=4)\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 498
        },
        "id": "8SO1C-4BP8iZ",
        "outputId": "3f86fbd3-5e16-40da-e660-2e47760c79f1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1008x504 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "kkxoHpzHPM4H"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KgwZLuM0J2_u"
      },
      "source": [
        "### Problem 11: Improving the Model\n",
        "\n",
        "Now that we have some basic models on the board, we want to try to improve these.  Below, we list a few things to explore in this pursuit.\n",
        "\n",
        "- More feature engineering and exploration.  For example, should we keep the gender feature?  Why or why not?\n",
        "- Hyperparameter tuning and grid search.  All of our models have additional hyperparameters to tune and explore.  For example the number of neighbors in KNN or the maximum depth of a Decision Tree.  \n",
        "- Adjust your performance metric"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model Optimization\n"
      ],
      "metadata": {
        "id": "L78sNTqqi9Z7"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Logistic Regression"
      ],
      "metadata": {
        "id": "PLFu2RbtjmX0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=42)"
      ],
      "metadata": {
        "id": "zbkF0UcqkZ93"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kFsAaViZJ2_v",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4bf75bee-23a9-41f6-a985-3d378117fd0b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:993: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n"
          ]
        }
      ],
      "source": [
        "logreg = LogisticRegression(random_state=42, class_weight='balanced').fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        ""
      ],
      "metadata": {
        "id": "NJiIkNUUjlwg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import classification_report, plot_confusion_matrix, mean_squared_error\n",
        "from sklearn.model_selection import GridSearchCV"
      ],
      "metadata": {
        "id": "U0Vnq01xlIPr"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "iUMaiS99J2_v",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0475ee6a-6701-448e-8090-482f2fb3c87e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification report for balanced classes \n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.92      0.47      0.62      9094\n",
            "           1       0.15      0.71      0.25      1200\n",
            "\n",
            "    accuracy                           0.50     10294\n",
            "   macro avg       0.54      0.59      0.43     10294\n",
            "weighted avg       0.83      0.50      0.58     10294\n",
            "\n",
            "\n",
            "Run time for baseline model :  0.10044217109680176\n",
            "Train MSE: 0.5050191049802474\n",
            "Test MSE: 0.504954342335341\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning: Precision is ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ]
        }
      ],
      "source": [
        "start_time = time.time()##\n",
        "\n",
        "train_preds = logreg.predict(X_train)\n",
        "test_preds = logreg.predict(X_test)\n",
        "\n",
        "print('Classification report for balanced classes \\n',classification_report(y_test,test_preds))\n",
        "test_preds = np.where(test_preds == True, 1, 0)\n",
        "train_preds = np.where(train_preds == True, 1, 0)\n",
        "\n",
        "# Evaluate Model\n",
        "svc_eval = evaluate_model(nb, X_test, y_test)\n",
        "\n",
        "end_time = time.time()##\n",
        "\n",
        "print(\"\\nRun time for baseline model : \", end_time-start_time)\n",
        "test_mse = mean_squared_error(test_preds, y_test)\n",
        "train_mse = mean_squared_error(train_preds, y_train)\n",
        "print(\"Train MSE:\", train_mse)\n",
        "print(\"Test MSE:\", test_mse)\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "logreg.get_params()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "m8ZFFNtnAFJf",
        "outputId": "25433e8c-daef-437c-900c-5a728271b40c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'C': 1.0,\n",
              " 'class_weight': 'balanced',\n",
              " 'dual': False,\n",
              " 'fit_intercept': True,\n",
              " 'intercept_scaling': 1,\n",
              " 'l1_ratio': None,\n",
              " 'max_iter': 100,\n",
              " 'multi_class': 'auto',\n",
              " 'n_jobs': None,\n",
              " 'penalty': 'l2',\n",
              " 'random_state': 42,\n",
              " 'solver': 'lbfgs',\n",
              " 'tol': 0.0001,\n",
              " 'verbose': 0,\n",
              " 'warm_start': False}"
            ]
          },
          "metadata": {},
          "execution_count": 330
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_confusion_matrix(logreg,X_test,y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 356
        },
        "id": "0DfFzFQmiTQH",
        "outputId": "93c41451-54ef-4cc4-ab89-e2917ffdc9cd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/deprecation.py:87: FutureWarning: Function plot_confusion_matrix is deprecated; Function `plot_confusion_matrix` is deprecated in 1.0 and will be removed in 1.2. Use one of the class methods: ConfusionMatrixDisplay.from_predictions or ConfusionMatrixDisplay.from_estimator.\n",
            "  warnings.warn(msg, category=FutureWarning)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x7f5764ef0d10>"
            ]
          },
          "metadata": {},
          "execution_count": 331
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "logistic_score_test = logreg.score(X_test, y_test)\n",
        "print(\"Accuracy = \\t\", \"{:.3f}\".format(logistic_score_test*100))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-J-4Lly0iTWT",
        "outputId": "d54cc59d-b7b9-47a0-d8d0-483fd68b832e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy = \t 49.505\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "solvers = ['newton-cg', 'lbfgs', 'liblinear']\n",
        "solvers = ['lbfgs']\n",
        "\n",
        "penalty = ['l2']\n",
        "c_values = [10, 1.0, 0.1, 0.01]\n",
        "grid = dict(solver=solvers,penalty=penalty,C=c_values)\n",
        "\n",
        "logreg_grid = GridSearchCV(logreg, param_grid=grid, n_jobs=-1, scoring='accuracy',error_score=0).fit(X_train, y_train)\n",
        "logreg_cv_res = logreg_grid.cv_results_\n",
        "\n",
        "logreg_mean_fit_time = logreg_cv_res['mean_fit_time'][0]\n",
        "\n",
        "print(\"Mean Fit Time: \", logreg_mean_fit_time)\n",
        "\n",
        "logreg_grid_score_train = logreg_grid.score(X_train, y_train)\n",
        "logreg_grid_score_test = logreg_grid.score(X_test, y_test)\n",
        "print(\"LogReg Train Score:\", logreg_grid_score_train)\n",
        "print(\"LogReg Test Score:\", logreg_grid_score_test)\n",
        "print(\" \")\n",
        "\n",
        "logreg_acc = (500+92)/(500+209+33+92)*100\n",
        "logreg_prec = (92)/(209+92)*100\n",
        "logreg_rec = (92)/(33+92)*100\n",
        "logreg_spec = (500)/(500+209)*100\n",
        "\n",
        "print(\"Accuracy = \\t\", \"{:.10f}\".format(logreg_acc))\n",
        "print(\"Precision = \\t\", \"{:.10f}\".format(logreg_prec))\n",
        "print(\"Recall = \\t\", \"{:.10f}\".format(logreg_rec))\n",
        "print(\"Specificity = \\t\", \"{:.10f}\".format(logreg_spec))\n",
        "\n",
        "print(\"Train MSE:\", train_mse)\n",
        "print(\"Test MSE:\", test_mse)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rpoGnevQFaxW",
        "outputId": "3b5a6ff4-60d2-4a3d-9801-734b4c6eec6f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mean Fit Time:  0.09887723922729492\n",
            "LogReg Train Score: 0.4949808950197526\n",
            "LogReg Test Score: 0.495045657664659\n",
            " \n",
            "Accuracy = \t 70.9832134293\n",
            "Precision = \t 30.5647840532\n",
            "Recall = \t 73.6000000000\n",
            "Specificity = \t 70.5218617772\n",
            "Train MSE: 0.5050191049802474\n",
            "Test MSE: 0.504954342335341\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:993: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "epGexL29FcNA"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Decision Tree**\n"
      ],
      "metadata": {
        "id": "BK3n0v-wG3-K"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.tree import DecisionTreeClassifier\n",
        "dtree = DecisionTreeClassifier(random_state=42, class_weight='balanced').fit(X_train, y_train)"
      ],
      "metadata": {
        "id": "sbuzBjhDG4Wg"
      },
      "execution_count": 340,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "dtree_train_preds = dtree.predict(X_train)\n",
        "dtree_test_preds = dtree.predict(X_test)\n",
        "\n",
        "print('Classification report for balanced classes \\n',classification_report(y_test,dtree_test_preds))\n",
        "dtree_test_preds = np.where(dtree_test_preds == True, 1, 0)\n",
        "dtree_train_preds = np.where(dtree_train_preds == True, 1, 0)\n",
        "\n",
        "dtree_test_mse = mean_squared_error(dtree_test_preds, y_test)\n",
        "dtree_train_mse = mean_squared_error(dtree_train_preds, y_train)\n",
        "\n",
        "print(\"DTree Train MSE:\", dtree_train_mse)\n",
        "print(\"DTree Test MSE:\", dtree_test_mse)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eHFQu6TxG4ai",
        "outputId": "26f8190e-66ca-4ad3-f8c1-215c289c93cf"
      },
      "execution_count": 341,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification report for balanced classes \n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.92      0.58      0.71      9094\n",
            "           1       0.16      0.61      0.25      1200\n",
            "\n",
            "    accuracy                           0.58     10294\n",
            "   macro avg       0.54      0.59      0.48     10294\n",
            "weighted avg       0.83      0.58      0.66     10294\n",
            "\n",
            "DTree Train MSE: 0.3821319862703193\n",
            "DTree Test MSE: 0.41703905187487855\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dtree.get_params()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_EKRMM4tG4gR",
        "outputId": "8c6cfc7f-7cf8-4f2e-b7f1-c7bb5e6e8688"
      },
      "execution_count": 342,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'ccp_alpha': 0.0,\n",
              " 'class_weight': 'balanced',\n",
              " 'criterion': 'gini',\n",
              " 'max_depth': None,\n",
              " 'max_features': None,\n",
              " 'max_leaf_nodes': None,\n",
              " 'min_impurity_decrease': 0.0,\n",
              " 'min_samples_leaf': 1,\n",
              " 'min_samples_split': 2,\n",
              " 'min_weight_fraction_leaf': 0.0,\n",
              " 'random_state': 42,\n",
              " 'splitter': 'best'}"
            ]
          },
          "metadata": {},
          "execution_count": 342
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_confusion_matrix(dtree,X_test,y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 356
        },
        "id": "OaPS-GPuHaFC",
        "outputId": "6d22e25c-c289-49a3-d5c1-449d53b598a4"
      },
      "execution_count": 343,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/deprecation.py:87: FutureWarning: Function plot_confusion_matrix is deprecated; Function `plot_confusion_matrix` is deprecated in 1.0 and will be removed in 1.2. Use one of the class methods: ConfusionMatrixDisplay.from_predictions or ConfusionMatrixDisplay.from_estimator.\n",
            "  warnings.warn(msg, category=FutureWarning)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x7f576538a1d0>"
            ]
          },
          "metadata": {},
          "execution_count": 343
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "splitter = ['best']\n",
        "\n",
        "criterion = ['gini']\n",
        "ccp_alphas = [10, 1.0, 0.1, 0.01]\n",
        "grid = dict(splitter=splitter,criterion=criterion,ccp_alpha=ccp_alphas)\n",
        "\n",
        "dtree_grid = GridSearchCV(dtree, param_grid=grid, n_jobs=-1, scoring='accuracy',error_score=0).fit(X_train, y_train)\n",
        "dtree_cv_res = dtree_grid.cv_results_\n",
        "\n",
        "dtree_mean_fit_time = dtree_cv_res['mean_fit_time'][0]\n",
        "\n",
        "print(\"Mean Fit Time: \", dtree_mean_fit_time)\n",
        "\n",
        "dtree_grid_score_train = dtree_grid.score(X_train, y_train)\n",
        "dtree_grid_score_test = dtree_grid.score(X_test, y_test)\n",
        "print(\"LogReg Train Score:\", dtree_grid_score_train)\n",
        "print(\"LogReg Test Score:\", dtree_grid_score_test)\n",
        "print(\" \")\n",
        "\n",
        "dtree_acc = (648+71)/(648+61+54+71)*100\n",
        "dtree_pre = (71)/(71+61)*100\n",
        "dtree_rec = (71)/(71+54)*100\n",
        "dtree_spe = (648)/(648+61)*100\n",
        "\n",
        "\n",
        "print(\"Accuracy = \\t\", \"{:.10f}\".format(dtree_acc))\n",
        "print(\"Precision = \\t\", \"{:.10f}\".format(dtree_pre))\n",
        "print(\"Recall = \\t\", \"{:.10f}\".format(dtree_rec))\n",
        "print(\"Specificity = \\t\", \"{:.10f}\".format(dtree_spe))\n",
        "\n",
        "print(\"Train MSE:\", dtree_train_mse)\n",
        "print(\"Test MSE:\", dtree_test_mse)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RIy4J8YvHaKT",
        "outputId": "a74c611d-0454-434e-bbaf-816828abb2ad"
      },
      "execution_count": 344,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mean Fit Time:  0.14212250709533691\n",
            "LogReg Train Score: 0.2981995984716016\n",
            "LogReg Test Score: 0.30532348941130755\n",
            " \n",
            "Accuracy = \t 86.2110311751\n",
            "Precision = \t 53.7878787879\n",
            "Recall = \t 56.8000000000\n",
            "Specificity = \t 91.3963328632\n",
            "Train MSE: 0.3821319862703193\n",
            "Test MSE: 0.41703905187487855\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "-qilk_2bHaOd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# KNN Classifier"
      ],
      "metadata": {
        "id": "d0JOS9IPHsYB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "knn = KNeighborsClassifier().fit(X_train, y_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GfjFi8gVHaVK",
        "outputId": "7adf57d3-6a68-4dd9-85b6-64d01a7c542a"
      },
      "execution_count": 351,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/neighbors/_classification.py:198: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
            "  return self._fit(X, y)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "knn_train_preds = knn.predict(X_train)\n",
        "knn_test_preds = knn.predict(X_test)\n",
        "\n",
        "print('Classification report for balanced classes \\n',classification_report(y_test,knn_test_preds))\n",
        "knn_test_preds = np.where(knn_test_preds == True, 1, 0)\n",
        "knn_train_preds = np.where(knn_train_preds == True, 1, 0)\n",
        "\n",
        "knn_test_mse = mean_squared_error(knn_test_preds, y_test)\n",
        "knn_train_mse = mean_squared_error(knn_train_preds, y_train)\n",
        "\n",
        "print(\"KNN Train MSE:\", knn_train_mse)\n",
        "print(\"KNN Test MSE:\", knn_test_mse)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "P0t3CWtIKCl-",
        "outputId": "83101fe7-80c4-4113-b65c-cd31c0512a30"
      },
      "execution_count": 352,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification report for balanced classes \n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.89      0.99      0.93      9094\n",
            "           1       0.33      0.05      0.09      1200\n",
            "\n",
            "    accuracy                           0.88     10294\n",
            "   macro avg       0.61      0.52      0.51     10294\n",
            "weighted avg       0.82      0.88      0.84     10294\n",
            "\n",
            "KNN Train MSE: 0.11556893983550288\n",
            "KNN Test MSE: 0.12298426267728774\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "knn.get_params()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tIRY6tjXKE4X",
        "outputId": "8df46088-561d-4cbe-8c93-2c98462ee766"
      },
      "execution_count": 353,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'algorithm': 'auto',\n",
              " 'leaf_size': 30,\n",
              " 'metric': 'minkowski',\n",
              " 'metric_params': None,\n",
              " 'n_jobs': None,\n",
              " 'n_neighbors': 5,\n",
              " 'p': 2,\n",
              " 'weights': 'uniform'}"
            ]
          },
          "metadata": {},
          "execution_count": 353
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_confusion_matrix(knn,X_test,y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 356
        },
        "id": "Dus7hmtoKE9R",
        "outputId": "904e2b53-b8c7-4336-dd9d-ff46fbf7143b"
      },
      "execution_count": 354,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/deprecation.py:87: FutureWarning: Function plot_confusion_matrix is deprecated; Function `plot_confusion_matrix` is deprecated in 1.0 and will be removed in 1.2. Use one of the class methods: ConfusionMatrixDisplay.from_predictions or ConfusionMatrixDisplay.from_estimator.\n",
            "  warnings.warn(msg, category=FutureWarning)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x7f5765e559d0>"
            ]
          },
          "metadata": {},
          "execution_count": 354
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "leaf_size = [30,20,10]\n",
        "\n",
        "metric = ['minkowski']\n",
        "n_neighbors = [10, 5, 3, 2]\n",
        "grid = dict(leaf_size=leaf_size,metric=metric,n_neighbors=n_neighbors)\n",
        "\n",
        "knn_grid = GridSearchCV(knn, param_grid=grid, n_jobs=-1, scoring='accuracy',error_score=0).fit(X_train, y_train)\n",
        "knn_cv_res = knn_grid.cv_results_\n",
        "\n",
        "knn_mean_fit_time = knn_cv_res['mean_fit_time'][0]\n",
        "print(\"Mean Fit Time: \", knn_mean_fit_time)\n",
        "\n",
        "knn_grid_score_train = knn_grid.score(X_train, y_train)\n",
        "knn_grid_score_test = knn_grid.score(X_test, y_test)\n",
        "print(\"LogReg Train Score:\", dtree_grid_score_train)\n",
        "print(\"LogReg Test Score:\", dtree_grid_score_test)\n",
        "print(\" \")\n",
        "\n",
        "knn_acc = (8963+65)/(8963+131+1135+65)*100\n",
        "knn_pre = (65)/(65+131)*100\n",
        "knn_rec = (65)/(65+1135)*100\n",
        "knn_spe = (8963)/(8963+131)*100\n",
        "\n",
        "\n",
        "print(\"Accuracy = \\t\", \"{:.10f}\".format(knn_acc))\n",
        "print(\"Precision = \\t\", \"{:.10f}\".format(knn_pre))\n",
        "print(\"Recall = \\t\", \"{:.10f}\".format(knn_rec))\n",
        "print(\"Specificity = \\t\", \"{:.10f}\".format(knn_spe))\n",
        "\n",
        "print(\"Train MSE:\", knn_train_mse)\n",
        "print(\"Test MSE:\", knn_test_mse)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ye9R_LOlH49S",
        "outputId": "c01cb219-1cac-44cc-9a36-882b13369eca"
      },
      "execution_count": 355,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/neighbors/_classification.py:198: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
            "  return self._fit(X, y)\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mean Fit Time:  0.04938406944274902\n",
            "LogReg Train Score: 0.2981995984716016\n",
            "LogReg Test Score: 0.30532348941130755\n",
            " \n",
            "Accuracy = \t 87.7015737323\n",
            "Precision = \t 33.1632653061\n",
            "Recall = \t 5.4166666667\n",
            "Specificity = \t 98.5594897735\n",
            "Train MSE: 0.11556893983550288\n",
            "Test MSE: 0.12298426267728774\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "xLYFJoAQH5BQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Support Vector Machine (SVC)"
      ],
      "metadata": {
        "id": "WJQs3Gq9IInz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.svm import SVC\n",
        "svc = SVC(class_weight='balanced').fit(X_train, y_train)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AGpR3yXmH5FY",
        "outputId": "a37cf7a2-a6d7-4138-8050-13cd587e73c5"
      },
      "execution_count": 358,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:993: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "svc_train_preds = svc.predict(X_train)\n",
        "svc_test_preds = svc.predict(X_test)\n",
        "\n",
        "print('Classification report for balanced classes \\n',classification_report(y_test,svc_test_preds))\n",
        "svc_test_preds = np.where(svc_test_preds == True, 1, 0)\n",
        "svc_train_preds = np.where(svc_train_preds == True, 1, 0)\n",
        "\n",
        "svc_test_mse = mean_squared_error(svc_test_preds, y_test)\n",
        "svc_train_mse = mean_squared_error(svc_train_preds, y_train)\n",
        "\n",
        "print(\"KNN Train MSE:\", svc_train_mse)\n",
        "print(\"KNN Test MSE:\", svc_test_mse)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Jkn2lWBuKsoQ",
        "outputId": "bc8c1f8d-6458-4da7-edb8-e4dd8a0063c5"
      },
      "execution_count": 359,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification report for balanced classes \n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.93      0.53      0.67      9094\n",
            "           1       0.16      0.69      0.26      1200\n",
            "\n",
            "    accuracy                           0.54     10294\n",
            "   macro avg       0.54      0.61      0.47     10294\n",
            "weighted avg       0.84      0.54      0.62     10294\n",
            "\n",
            "KNN Train MSE: 0.4599766854478337\n",
            "KNN Test MSE: 0.4553137750145716\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "svc.get_params()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "X7uY9sZXKsse",
        "outputId": "56800c5e-60ef-42e5-a0fc-4081037711fd"
      },
      "execution_count": 360,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'C': 1.0,\n",
              " 'break_ties': False,\n",
              " 'cache_size': 200,\n",
              " 'class_weight': 'balanced',\n",
              " 'coef0': 0.0,\n",
              " 'decision_function_shape': 'ovr',\n",
              " 'degree': 3,\n",
              " 'gamma': 'scale',\n",
              " 'kernel': 'rbf',\n",
              " 'max_iter': -1,\n",
              " 'probability': False,\n",
              " 'random_state': None,\n",
              " 'shrinking': True,\n",
              " 'tol': 0.001,\n",
              " 'verbose': False}"
            ]
          },
          "metadata": {},
          "execution_count": 360
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_confusion_matrix(svc,X_test,y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 356
        },
        "id": "EPR4Rj7xKswd",
        "outputId": "b5ce6a6c-ef9d-4ec6-9f73-8a78c9a01375"
      },
      "execution_count": 361,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/deprecation.py:87: FutureWarning: Function plot_confusion_matrix is deprecated; Function `plot_confusion_matrix` is deprecated in 1.0 and will be removed in 1.2. Use one of the class methods: ConfusionMatrixDisplay.from_predictions or ConfusionMatrixDisplay.from_estimator.\n",
            "  warnings.warn(msg, category=FutureWarning)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x7f5767e49b10>"
            ]
          },
          "metadata": {},
          "execution_count": 361
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "kernel = ['rbf']\n",
        "\n",
        "decision_function_shape = ['ovr']\n",
        "c_values = [10, 1.0, 0.1, 0.01]\n",
        "grid = dict(kernel=kernel,decision_function_shape=decision_function_shape,C=c_values)\n",
        "\n",
        "svc_grid = GridSearchCV(svc, param_grid=grid, n_jobs=-1, scoring='accuracy',error_score=0).fit(X_train, y_train)\n",
        "svc_cv_res = svc_grid.cv_results_\n",
        "\n",
        "svc_mean_fit_time = svc_cv_res['mean_fit_time'][0]\n",
        "print(\"Mean Fit Time: \", svc_mean_fit_time)\n",
        "\n",
        "svc_grid_score_train = svc_grid.score(X_train, y_train)\n",
        "svc_grid_score_test = svc_grid.score(X_test, y_test)\n",
        "print(\"LogReg Train Score:\", svc_grid_score_train)\n",
        "print(\"LogReg Test Score:\", svc_grid_score_test)\n",
        "print(\" \")\n",
        "\n",
        "svc_acc = (709+0)/(709+0+125+0)*100\n",
        "svc_pre = 0\n",
        "svc_rec = (0)/(0+125)*100\n",
        "svc_spe = (709)/(709+0)*100\n",
        "\n",
        "\n",
        "print(\"Accuracy = \\t\", \"{:.10f}\".format(svc_acc))\n",
        "print(\"Precision = \\t\", \"{:.10f}\".format(svc_pre))\n",
        "print(\"Recall = \\t\", \"{:.10f}\".format(svc_rec))\n",
        "print(\"Specificity = \\t\", \"{:.10f}\".format(svc_spe))\n",
        "\n",
        "print(\"Train MSE:\", svc_train_mse)\n",
        "print(\"Test MSE:\", svc_test_mse)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9tn-5u8zKs3a",
        "outputId": "4432d598-1232-47c1-cd0c-db2abdd1ae17"
      },
      "execution_count": 362,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/validation.py:993: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
            "  y = column_or_1d(y, warn=True)\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mean Fit Time:  87.15264172554016\n",
            "LogReg Train Score: 0.5561168318114111\n",
            "LogReg Test Score: 0.5533320380804352\n",
            " \n",
            "Accuracy = \t 85.0119904077\n",
            "Precision = \t 0.0000000000\n",
            "Recall = \t 0.0000000000\n",
            "Specificity = \t 100.0000000000\n",
            "Train MSE: 0.4599766854478337\n",
            "Test MSE: 0.4553137750145716\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Classifiers Comparison Table\n",
        "\n"
      ],
      "metadata": {
        "id": "G0vG_aMhKo5B"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "val_df = pd.DataFrame(columns = ['model','train score','test score','Mean fit time', 'Accuracy', 'Precision', 'Recall', 'Specificity', 'Train MSE', 'Test MSE'])"
      ],
      "metadata": {
        "id": "Odb-U56yKs73"
      },
      "execution_count": 363,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "val_df.loc[len(val_df.index)] = ['Logistic Regression', logreg_grid_score_train, logreg_grid_score_test, logreg_mean_fit_time, logreg_acc, logreg_pre, logreg_rec, logreg_spe, train_mse, test_mse] \n",
        "val_df.loc[len(val_df.index)] = ['Decision Tree', dtree_grid_score_train, dtree_grid_score_test, dtree_mean_fit_time, dtree_acc, dtree_pre, dtree_rec, dtree_spe, dtree_train_mse, dtree_test_mse]\n",
        "val_df.loc[len(val_df.index)] = ['KNearestNeighbors', knn_grid_score_train, knn_grid_score_test, knn_mean_fit_time, knn_acc, knn_pre, knn_rec, knn_spe, knn_train_mse, knn_test_mse]\n",
        "val_df.loc[len(val_df.index)] = ['Support Vector Machine', svc_grid_score_train, svc_grid_score_test, svc_mean_fit_time, svc_acc, svc_pre, svc_rec, svc_spe, svc_train_mse, svc_test_mse]"
      ],
      "metadata": {
        "id": "rZTus1ayKtAr"
      },
      "execution_count": 364,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "val_df"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 173
        },
        "id": "SWTrH7F9KtG7",
        "outputId": "e8e56d39-4ee5-4a12-d3bb-47c0822c14b5"
      },
      "execution_count": 365,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                    model  train score  test score  Mean fit time   Accuracy  \\\n",
              "0     Logistic Regression     0.494981    0.495046       0.098877  70.983213   \n",
              "1           Decision Tree     0.298200    0.305323       0.142123  86.211031   \n",
              "2       KNearestNeighbors     0.886147    0.882067       0.049384  87.701574   \n",
              "3  Support Vector Machine     0.556117    0.553332      87.152642  85.011990   \n",
              "\n",
              "   Precision     Recall  Specificity  Train MSE  Test MSE  \n",
              "0  14.942126  73.600000    46.668133   0.505019  0.504954  \n",
              "1  53.787879  56.800000    91.396333   0.382132  0.417039  \n",
              "2  33.163265   5.416667    98.559490   0.115569  0.122984  \n",
              "3   0.000000   0.000000   100.000000   0.459977  0.455314  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-2c4924b7-92be-4602-84d5-8cbbf77f3ed9\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>model</th>\n",
              "      <th>train score</th>\n",
              "      <th>test score</th>\n",
              "      <th>Mean fit time</th>\n",
              "      <th>Accuracy</th>\n",
              "      <th>Precision</th>\n",
              "      <th>Recall</th>\n",
              "      <th>Specificity</th>\n",
              "      <th>Train MSE</th>\n",
              "      <th>Test MSE</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Logistic Regression</td>\n",
              "      <td>0.494981</td>\n",
              "      <td>0.495046</td>\n",
              "      <td>0.098877</td>\n",
              "      <td>70.983213</td>\n",
              "      <td>14.942126</td>\n",
              "      <td>73.600000</td>\n",
              "      <td>46.668133</td>\n",
              "      <td>0.505019</td>\n",
              "      <td>0.504954</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Decision Tree</td>\n",
              "      <td>0.298200</td>\n",
              "      <td>0.305323</td>\n",
              "      <td>0.142123</td>\n",
              "      <td>86.211031</td>\n",
              "      <td>53.787879</td>\n",
              "      <td>56.800000</td>\n",
              "      <td>91.396333</td>\n",
              "      <td>0.382132</td>\n",
              "      <td>0.417039</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>KNearestNeighbors</td>\n",
              "      <td>0.886147</td>\n",
              "      <td>0.882067</td>\n",
              "      <td>0.049384</td>\n",
              "      <td>87.701574</td>\n",
              "      <td>33.163265</td>\n",
              "      <td>5.416667</td>\n",
              "      <td>98.559490</td>\n",
              "      <td>0.115569</td>\n",
              "      <td>0.122984</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Support Vector Machine</td>\n",
              "      <td>0.556117</td>\n",
              "      <td>0.553332</td>\n",
              "      <td>87.152642</td>\n",
              "      <td>85.011990</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>100.000000</td>\n",
              "      <td>0.459977</td>\n",
              "      <td>0.455314</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-2c4924b7-92be-4602-84d5-8cbbf77f3ed9')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-2c4924b7-92be-4602-84d5-8cbbf77f3ed9 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-2c4924b7-92be-4602-84d5-8cbbf77f3ed9');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 365
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "6gK7dLElG4m1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Analysis Report\n",
        "\n",
        "The Analysis of the Logistic Regress, Decision Tree Classifier, KNeighborsClassifier and the Support Vector Machines was performed according to the following criteria\n",
        "1. Imbalance Class Handling\n",
        "2. Model Training Speed\n",
        "3. Interpretable Results\n",
        "\n",
        "Other criteria observed include\n",
        "1. Accuracy\n",
        "2. Precision\n",
        "3. Recall\n",
        "4. Specificity\n",
        "5. Mean Squared Error"
      ],
      "metadata": {
        "id": "Yeu8ikKVRNUz"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "- Logistic Regression Classifier\n",
        "  - Handles Imbalanced Classes\n",
        "  - Speed of Training is moderately high at a Mean Fit Time of 0.09\n",
        "  - Test Score was low at 50%, close to the same as the Train Score at 49%\n",
        "  - Accuracy and Specificity are not too high at 71%.\n",
        "  - Precision is low at 14.9%\n",
        "  - Recall is high at 73%\n",
        "  - It appears not to overfit\n",
        "\n",
        "- Decision Tree Classifier\n",
        "  - Handles Imbalanced Classes\n",
        "  - Speed of Training is moderately high at a Mean Fit Time of 0.14\n",
        "  - Accuracy and Specificity are very high at 86% and 91%, respectively\n",
        "  - The Test Score performs slightly better than the Train Score\n",
        "  - Precision is above average at 53%\n",
        "  - Recall is above average at 57%\n",
        "  - However, the Decision Tree Classifier appears to overfit with Test MSE higher than Train MSE\n",
        "\n",
        "- KNearest Neighbors Classifier (KNN)\n",
        "  - KNN cannot handle Imbalanced classes\n",
        "  - Speed of Training is very high at a Mean Fit Time of 0.04\n",
        "  - Accuracy and Specificity are very high at 86% and 98%, respectively\n",
        "  - The Train Score performs slightly better than the Test Score\n",
        "  - Precision is low at 33%\n",
        "  - Recall is very low at 5.4%\n",
        "  - However, the KNNeighbors Classifier appears to overfit with Test MSE higher than Train MSE\n",
        "   \n",
        "- Support Vector Machine\n",
        "  - SVM handles Imbalanced classes\n",
        "  - Speed of Training was slow with a Mean Fit Time of 87.2\n",
        "  - Accuracy and Specificity are very high at 85% and 100%, respectively\n",
        "  - Precision is at 0\n",
        "  - Recall is low at 0%\n",
        "  - However, the KNNeighbors Classifier appears to not overfit with \t\n",
        "  - Test MSE and Train MSE are close at 45%\n",
        "- The Test Score performs better than the Train Score\n",
        "\n",
        "## Selection: Logistic Regression\n",
        "- The Logistic Regression Classifier is the one I would select based on the above table\n",
        "- Decision Tree is my second model of choice after Logistic Regression"
      ],
      "metadata": {
        "id": "z1quFWYtRNbC"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Conclusion\n",
        "\n",
        "For a simple model we can see that our model did decently on classifying the data. But there are still some weakness on our model, especially shown on the recall metric where we only get about 73%. This means that our model are only able to detect 73% of potential customer and miss the other 27%. The result is not that much different after optimising the model using `GridSearchCV` which can means that we hit our limit with this model. To improve our performance we can try to look into another algorithm such as `GradientBoostingClassifier`.\n",
        "\n",
        "# References\n",
        "\n",
        "1. [Scikit-learn Documentation](https://scikit-learn.org/stable/index.html)\n",
        "2. [The 5 Classification Evaluation metrics every Data Scientist must know](https://towardsdatascience.com/the-5-classification-evaluation-metrics-you-must-know-aa97784ff226)\n",
        "3. [The Python Graph Gallery - Grouped Bar Plot](https://python-graph-gallery.com/11-grouped-barplot/)"
      ],
      "metadata": {
        "id": "1KNn6hBqiVyO"
      }
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.8.8"
    },
    "colab": {
      "name": "Practical Application III-TB-1-FeatureEngr-Modelling-Classification-Final.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}