FML_Assignment.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "FML_Assignment.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Gr1tBSyOb-cP",
        "outputId": "059c78a3-d24d-412f-a848-9a55df502a0a"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive/')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive/\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ocVmOJGqdQ0U"
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.decomposition import PCA\n",
        "from sklearn.impute import SimpleImputer\n",
        "from sklearn.feature_selection import RFECV\n",
        "%matplotlib inline\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import RepeatedStratifiedKFold\n",
        "from numpy import mean\n",
        "from numpy import std\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.model_selection import learning_curve\n",
        "from sklearn.feature_selection import SelectKBest\n",
        "# classifiers\n",
        "from sklearn.naive_bayes import GaussianNB\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis\n",
        "from sklearn.svm import SVC\n",
        "from sklearn.neural_network import MLPClassifier"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TalIAk3hcSvX"
      },
      "source": [
        "# load data\n",
        "training1=pd.read_csv('/content/drive/My Drive/FML_Assignment/training1.csv')\n",
        "training2=pd.read_csv('/content/drive/My Drive/FML_Assignment/training2.csv')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "31jxXe9Bypei",
        "outputId": "486a15d3-d7cd-4c4d-815b-bc7daae185d7"
      },
      "source": [
        "dups = data.duplicated()\n",
        "# identify that contain duplicate observations.\n",
        "print(dups.any())\n",
        "print(data[dups])\n",
        "print('____________')\n",
        "# identify column variables that only have a single value.\n",
        "counts = data.nunique()\n",
        "to_del = [i for i,v in enumerate(counts) if v == 1]\n",
        "print(to_del)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "False\n",
            "Empty DataFrame\n",
            "Columns: [CNNs, CNNs.1, CNNs.2, CNNs.3, CNNs.4, CNNs.5, CNNs.6, CNNs.7, CNNs.8, CNNs.9, CNNs.10, CNNs.11, CNNs.12, CNNs.13, CNNs.14, CNNs.15, CNNs.16, CNNs.17, CNNs.18, CNNs.19, CNNs.20, CNNs.21, CNNs.22, CNNs.23, CNNs.24, CNNs.25, CNNs.26, CNNs.27, CNNs.28, CNNs.29, CNNs.30, CNNs.31, CNNs.32, CNNs.33, CNNs.34, CNNs.35, CNNs.36, CNNs.37, CNNs.38, CNNs.39, CNNs.40, CNNs.41, CNNs.42, CNNs.43, CNNs.44, CNNs.45, CNNs.46, CNNs.47, CNNs.48, CNNs.49, CNNs.50, CNNs.51, CNNs.52, CNNs.53, CNNs.54, CNNs.55, CNNs.56, CNNs.57, CNNs.58, CNNs.59, CNNs.60, CNNs.61, CNNs.62, CNNs.63, CNNs.64, CNNs.65, CNNs.66, CNNs.67, CNNs.68, CNNs.69, CNNs.70, CNNs.71, CNNs.72, CNNs.73, CNNs.74, CNNs.75, CNNs.76, CNNs.77, CNNs.78, CNNs.79, CNNs.80, CNNs.81, CNNs.82, CNNs.83, CNNs.84, CNNs.85, CNNs.86, CNNs.87, CNNs.88, CNNs.89, CNNs.90, CNNs.91, CNNs.92, CNNs.93, CNNs.94, CNNs.95, CNNs.96, CNNs.97, CNNs.98, CNNs.99, ...]\n",
            "Index: []\n",
            "\n",
            "[0 rows x 4610 columns]\n",
            "____________\n",
            "[]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 425
        },
        "id": "4koia2rIc0az",
        "outputId": "be1b340e-e31e-4056-8263-834e52c2f66b"
      },
      "source": [
        "data=training2\n",
        "\n",
        "# drop the low confidence data\n",
        "data.drop(data.loc[data['confidence']==0.66].index, inplace=True)\n",
        "data"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>CNNs</th>\n",
              "      <th>CNNs.1</th>\n",
              "      <th>CNNs.2</th>\n",
              "      <th>CNNs.3</th>\n",
              "      <th>CNNs.4</th>\n",
              "      <th>CNNs.5</th>\n",
              "      <th>CNNs.6</th>\n",
              "      <th>CNNs.7</th>\n",
              "      <th>CNNs.8</th>\n",
              "      <th>CNNs.9</th>\n",
              "      <th>CNNs.10</th>\n",
              "      <th>CNNs.11</th>\n",
              "      <th>CNNs.12</th>\n",
              "      <th>CNNs.13</th>\n",
              "      <th>CNNs.14</th>\n",
              "      <th>CNNs.15</th>\n",
              "      <th>CNNs.16</th>\n",
              "      <th>CNNs.17</th>\n",
              "      <th>CNNs.18</th>\n",
              "      <th>CNNs.19</th>\n",
              "      <th>CNNs.20</th>\n",
              "      <th>CNNs.21</th>\n",
              "      <th>CNNs.22</th>\n",
              "      <th>CNNs.23</th>\n",
              "      <th>CNNs.24</th>\n",
              "      <th>CNNs.25</th>\n",
              "      <th>CNNs.26</th>\n",
              "      <th>CNNs.27</th>\n",
              "      <th>CNNs.28</th>\n",
              "      <th>CNNs.29</th>\n",
              "      <th>CNNs.30</th>\n",
              "      <th>CNNs.31</th>\n",
              "      <th>CNNs.32</th>\n",
              "      <th>CNNs.33</th>\n",
              "      <th>CNNs.34</th>\n",
              "      <th>CNNs.35</th>\n",
              "      <th>CNNs.36</th>\n",
              "      <th>CNNs.37</th>\n",
              "      <th>CNNs.38</th>\n",
              "      <th>CNNs.39</th>\n",
              "      <th>...</th>\n",
              "      <th>GIST.474</th>\n",
              "      <th>GIST.475</th>\n",
              "      <th>GIST.476</th>\n",
              "      <th>GIST.477</th>\n",
              "      <th>GIST.478</th>\n",
              "      <th>GIST.479</th>\n",
              "      <th>GIST.480</th>\n",
              "      <th>GIST.481</th>\n",
              "      <th>GIST.482</th>\n",
              "      <th>GIST.483</th>\n",
              "      <th>GIST.484</th>\n",
              "      <th>GIST.485</th>\n",
              "      <th>GIST.486</th>\n",
              "      <th>GIST.487</th>\n",
              "      <th>GIST.488</th>\n",
              "      <th>GIST.489</th>\n",
              "      <th>GIST.490</th>\n",
              "      <th>GIST.491</th>\n",
              "      <th>GIST.492</th>\n",
              "      <th>GIST.493</th>\n",
              "      <th>GIST.494</th>\n",
              "      <th>GIST.495</th>\n",
              "      <th>GIST.496</th>\n",
              "      <th>GIST.497</th>\n",
              "      <th>GIST.498</th>\n",
              "      <th>GIST.499</th>\n",
              "      <th>GIST.500</th>\n",
              "      <th>GIST.501</th>\n",
              "      <th>GIST.502</th>\n",
              "      <th>GIST.503</th>\n",
              "      <th>GIST.504</th>\n",
              "      <th>GIST.505</th>\n",
              "      <th>GIST.506</th>\n",
              "      <th>GIST.507</th>\n",
              "      <th>GIST.508</th>\n",
              "      <th>GIST.509</th>\n",
              "      <th>GIST.510</th>\n",
              "      <th>GIST.511</th>\n",
              "      <th>label</th>\n",
              "      <th>confidence</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>1.5151</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.46469</td>\n",
              "      <td>0.89557</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.41339</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>1.177400</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.34667</td>\n",
              "      <td>0.22934</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.61443</td>\n",
              "      <td>0.62352</td>\n",
              "      <td>0.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.043887</td>\n",
              "      <td>0.000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.092796</td>\n",
              "      <td>0.51648</td>\n",
              "      <td>0.66142</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.005754</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.039100</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.017372</td>\n",
              "      <td>0.007980</td>\n",
              "      <td>0.025694</td>\n",
              "      <td>0.009572</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.005846</td>\n",
              "      <td>0.010710</td>\n",
              "      <td>0.016408</td>\n",
              "      <td>0.011204</td>\n",
              "      <td>0.001122</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.006253</td>\n",
              "      <td>0.018301</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.056184</td>\n",
              "      <td>0.061991</td>\n",
              "      <td>0.020518</td>\n",
              "      <td>0.007175</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.054380</td>\n",
              "      <td>0.016573</td>\n",
              "      <td>0.003284</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.011694</td>\n",
              "      <td>0.002099</td>\n",
              "      <td>0.012144</td>\n",
              "      <td>0.001809</td>\n",
              "      <td>0.008753</td>\n",
              "      <td>0.004012</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.013421</td>\n",
              "      <td>0.005070</td>\n",
              "      <td>0.036088</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.29005</td>\n",
              "      <td>0.377550</td>\n",
              "      <td>0.172170</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.66253</td>\n",
              "      <td>1.30180</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.823700</td>\n",
              "      <td>0.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>1.1019</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.14064</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.67840</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>1.04680</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.306070</td>\n",
              "      <td>0.63127</td>\n",
              "      <td>0.89558</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>...</td>\n",
              "      <td>0.020151</td>\n",
              "      <td>0.017083</td>\n",
              "      <td>0.010079</td>\n",
              "      <td>0.014837</td>\n",
              "      <td>0.014841</td>\n",
              "      <td>0.028490</td>\n",
              "      <td>0.011267</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.006051</td>\n",
              "      <td>0.009580</td>\n",
              "      <td>0.006189</td>\n",
              "      <td>0.015256</td>\n",
              "      <td>0.003178</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.004871</td>\n",
              "      <td>0.005753</td>\n",
              "      <td>0.039314</td>\n",
              "      <td>0.021291</td>\n",
              "      <td>0.001760</td>\n",
              "      <td>0.061693</td>\n",
              "      <td>0.022594</td>\n",
              "      <td>0.023751</td>\n",
              "      <td>0.012351</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.041376</td>\n",
              "      <td>0.008531</td>\n",
              "      <td>0.011940</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.018484</td>\n",
              "      <td>0.005399</td>\n",
              "      <td>0.007041</td>\n",
              "      <td>0.007484</td>\n",
              "      <td>0.012602</td>\n",
              "      <td>0.002727</td>\n",
              "      <td>0.015594</td>\n",
              "      <td>0.012372</td>\n",
              "      <td>0.009573</td>\n",
              "      <td>0.008190</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28</th>\n",
              "      <td>0.98552</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000932</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>1.009100</td>\n",
              "      <td>0.000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>2.15140</td>\n",
              "      <td>...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.033568</td>\n",
              "      <td>0.006099</td>\n",
              "      <td>0.017287</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.058670</td>\n",
              "      <td>0.012437</td>\n",
              "      <td>0.033040</td>\n",
              "      <td>0.007975</td>\n",
              "      <td>0.007391</td>\n",
              "      <td>0.007200</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.015862</td>\n",
              "      <td>0.011259</td>\n",
              "      <td>0.007914</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.023108</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.040738</td>\n",
              "      <td>0.024548</td>\n",
              "      <td>0.050125</td>\n",
              "      <td>0.028176</td>\n",
              "      <td>0.029166</td>\n",
              "      <td>0.041450</td>\n",
              "      <td>0.010401</td>\n",
              "      <td>0.007129</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.017803</td>\n",
              "      <td>0.005348</td>\n",
              "      <td>0.005875</td>\n",
              "      <td>0.006137</td>\n",
              "      <td>0.009345</td>\n",
              "      <td>0.003796</td>\n",
              "      <td>0.038098</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.024351</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>36</th>\n",
              "      <td>0.34266</td>\n",
              "      <td>0.44640</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.406500</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.76886</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.12262</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.51334</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.68515</td>\n",
              "      <td>0.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.15730</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.031969</td>\n",
              "      <td>0.024067</td>\n",
              "      <td>0.031279</td>\n",
              "      <td>0.040002</td>\n",
              "      <td>0.014832</td>\n",
              "      <td>0.027478</td>\n",
              "      <td>0.025730</td>\n",
              "      <td>0.028632</td>\n",
              "      <td>0.021095</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.017050</td>\n",
              "      <td>0.017578</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.059241</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.025195</td>\n",
              "      <td>0.099179</td>\n",
              "      <td>0.025167</td>\n",
              "      <td>0.033253</td>\n",
              "      <td>0.033032</td>\n",
              "      <td>0.025688</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.030170</td>\n",
              "      <td>0.025373</td>\n",
              "      <td>0.044848</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.030316</td>\n",
              "      <td>0.026394</td>\n",
              "      <td>0.021039</td>\n",
              "      <td>0.034658</td>\n",
              "      <td>0.047409</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.015142</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.022586</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.028349</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>42</th>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.98419</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.62366</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.048117</td>\n",
              "      <td>0.059774</td>\n",
              "      <td>0.004919</td>\n",
              "      <td>0.010298</td>\n",
              "      <td>0.007150</td>\n",
              "      <td>0.018074</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.016881</td>\n",
              "      <td>0.007646</td>\n",
              "      <td>0.031194</td>\n",
              "      <td>0.032426</td>\n",
              "      <td>0.005368</td>\n",
              "      <td>0.018135</td>\n",
              "      <td>0.051592</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.022939</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.025380</td>\n",
              "      <td>0.014188</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.059960</td>\n",
              "      <td>0.049069</td>\n",
              "      <td>0.047452</td>\n",
              "      <td>0.053986</td>\n",
              "      <td>0.038580</td>\n",
              "      <td>0.004829</td>\n",
              "      <td>0.004814</td>\n",
              "      <td>0.037901</td>\n",
              "      <td>0.016311</td>\n",
              "      <td>0.003360</td>\n",
              "      <td>0.003254</td>\n",
              "      <td>0.035968</td>\n",
              "      <td>0.007106</td>\n",
              "      <td>0.053848</td>\n",
              "      <td>0.043616</td>\n",
              "      <td>0.045859</td>\n",
              "      <td>0.016011</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2978</th>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.358440</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.90717</td>\n",
              "      <td>0.80058</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.05752</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>2.40760</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>1.70130</td>\n",
              "      <td>0.47990</td>\n",
              "      <td>0.35644</td>\n",
              "      <td>0.554440</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.31996</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.257870</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.29765</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.473440</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.65331</td>\n",
              "      <td>...</td>\n",
              "      <td>0.066344</td>\n",
              "      <td>0.073337</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.028346</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.036594</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.037061</td>\n",
              "      <td>0.014526</td>\n",
              "      <td>0.019121</td>\n",
              "      <td>0.009981</td>\n",
              "      <td>0.048344</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.032792</td>\n",
              "      <td>0.010854</td>\n",
              "      <td>0.048209</td>\n",
              "      <td>0.020706</td>\n",
              "      <td>0.031312</td>\n",
              "      <td>0.025820</td>\n",
              "      <td>0.029370</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.029000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.045156</td>\n",
              "      <td>0.038654</td>\n",
              "      <td>0.016896</td>\n",
              "      <td>0.013499</td>\n",
              "      <td>0.029791</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.017124</td>\n",
              "      <td>0.015486</td>\n",
              "      <td>0.029014</td>\n",
              "      <td>0.011941</td>\n",
              "      <td>0.043056</td>\n",
              "      <td>0.015908</td>\n",
              "      <td>0.032228</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2987</th>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.511940</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.27121</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.13894</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.14461</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.027603</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.72122</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.53417</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.30583</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.43161</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000</td>\n",
              "      <td>1.33010</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.33036</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.035665</td>\n",
              "      <td>0.021756</td>\n",
              "      <td>0.018079</td>\n",
              "      <td>0.016238</td>\n",
              "      <td>0.015099</td>\n",
              "      <td>0.023200</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.023904</td>\n",
              "      <td>0.006413</td>\n",
              "      <td>0.020702</td>\n",
              "      <td>0.010731</td>\n",
              "      <td>0.011501</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.029195</td>\n",
              "      <td>0.004428</td>\n",
              "      <td>0.016763</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.040703</td>\n",
              "      <td>0.048912</td>\n",
              "      <td>0.030043</td>\n",
              "      <td>0.034479</td>\n",
              "      <td>0.055645</td>\n",
              "      <td>0.023252</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.034007</td>\n",
              "      <td>0.024211</td>\n",
              "      <td>0.018906</td>\n",
              "      <td>0.009078</td>\n",
              "      <td>0.009756</td>\n",
              "      <td>0.024363</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.039619</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.024016</td>\n",
              "      <td>0.070927</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2994</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.048602</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.11445</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.32577</td>\n",
              "      <td>0.70036</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.84342</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.22012</td>\n",
              "      <td>0.18668</td>\n",
              "      <td>0.37605</td>\n",
              "      <td>0.705730</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.12856</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.65742</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.252</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>...</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.051667</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.020125</td>\n",
              "      <td>0.045329</td>\n",
              "      <td>0.051442</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.042607</td>\n",
              "      <td>0.030276</td>\n",
              "      <td>0.034500</td>\n",
              "      <td>0.023245</td>\n",
              "      <td>0.041524</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.046234</td>\n",
              "      <td>0.026353</td>\n",
              "      <td>0.035897</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.032216</td>\n",
              "      <td>0.078904</td>\n",
              "      <td>0.060334</td>\n",
              "      <td>0.060547</td>\n",
              "      <td>0.029030</td>\n",
              "      <td>0.057119</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.027535</td>\n",
              "      <td>0.033689</td>\n",
              "      <td>0.014716</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.030388</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.030674</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.014262</td>\n",
              "      <td>0.028280</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.023965</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2995</th>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.385090</td>\n",
              "      <td>1.512700</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.58160</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.64208</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.72838</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.016287</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.5053</td>\n",
              "      <td>0.80474</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.70036</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.72068</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.6995</td>\n",
              "      <td>1.23110</td>\n",
              "      <td>...</td>\n",
              "      <td>0.017682</td>\n",
              "      <td>0.075365</td>\n",
              "      <td>0.011170</td>\n",
              "      <td>0.020524</td>\n",
              "      <td>0.004148</td>\n",
              "      <td>0.016940</td>\n",
              "      <td>0.015104</td>\n",
              "      <td>0.034050</td>\n",
              "      <td>0.011095</td>\n",
              "      <td>0.010095</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.005738</td>\n",
              "      <td>0.018486</td>\n",
              "      <td>0.036345</td>\n",
              "      <td>0.003406</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.037426</td>\n",
              "      <td>0.109120</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.022961</td>\n",
              "      <td>0.021211</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.037172</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.035435</td>\n",
              "      <td>0.023471</td>\n",
              "      <td>0.008260</td>\n",
              "      <td>0.026336</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.011719</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.004419</td>\n",
              "      <td>0.048982</td>\n",
              "      <td>0.016881</td>\n",
              "      <td>0.019765</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2997</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.639080</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.27171</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.35540</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.74870</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.670700</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.89880</td>\n",
              "      <td>...</td>\n",
              "      <td>0.041939</td>\n",
              "      <td>0.052368</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.017927</td>\n",
              "      <td>0.063527</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.043826</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.010698</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.019283</td>\n",
              "      <td>0.032178</td>\n",
              "      <td>0.026081</td>\n",
              "      <td>0.026755</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.043924</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.030764</td>\n",
              "      <td>0.047300</td>\n",
              "      <td>0.045167</td>\n",
              "      <td>0.040845</td>\n",
              "      <td>0.033875</td>\n",
              "      <td>0.035150</td>\n",
              "      <td>0.043500</td>\n",
              "      <td>0.033313</td>\n",
              "      <td>0.036923</td>\n",
              "      <td>0.014606</td>\n",
              "      <td>0.046039</td>\n",
              "      <td>0.014998</td>\n",
              "      <td>0.014979</td>\n",
              "      <td>0.029114</td>\n",
              "      <td>0.029143</td>\n",
              "      <td>0.012367</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.037802</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.063759</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>789 rows × 4610 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "         CNNs   CNNs.1    CNNs.2  ...  GIST.511  label  confidence\n",
              "14    0.00000  0.00000       NaN  ...  0.036088    0.0         1.0\n",
              "23        NaN  0.29005  0.377550  ...  0.008190    1.0         1.0\n",
              "28    0.98552      NaN  0.000000  ...       NaN    0.0         1.0\n",
              "36    0.34266  0.44640  0.000000  ...  0.028349    1.0         1.0\n",
              "42    0.00000  0.00000  0.000000  ...  0.016011    0.0         1.0\n",
              "...       ...      ...       ...  ...       ...    ...         ...\n",
              "2978  0.00000      NaN  0.000000  ...       NaN    1.0         1.0\n",
              "2987  0.00000      NaN  0.511940  ...  0.070927    0.0         1.0\n",
              "2994      NaN  0.00000  0.048602  ...  0.023965    1.0         1.0\n",
              "2995  0.00000  0.00000  0.385090  ...       NaN    0.0         1.0\n",
              "2997      NaN  0.00000  0.639080  ...  0.063759    1.0         1.0\n",
              "\n",
              "[789 rows x 4610 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YAPxHiEwggpG"
      },
      "source": [
        "# get data with label CNN and gist\n",
        "properties = list(data.columns.values)\n",
        "CNN_properties = [s for s in properties if \"CNNs.\" in s]\n",
        "CNN_data = data[CNN_properties]\n",
        "GIST_properties = [s for s in properties if \"GIST.\" in s]\n",
        "GIST_data = data[GIST_properties]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JnVHYfU3syoN"
      },
      "source": [
        "from xgboost import XGBClassifier\n",
        "from sklearn.model_selection import train_test_split\n",
        "# calculate the importance of CNN feature\n",
        "X_train, X_test, y_train, y_test = train_test_split(CNN_data, train_label, test_size=0.25, random_state=42)\n",
        "\n",
        "ss = MinMaxScaler(feature_range = (0,1))\n",
        "\n",
        "X_train_scaled = ss.fit_transform(X_train)\n",
        "X_test_scaled = ss.transform(X_test)\n",
        "\n",
        "model = XGBClassifier()\n",
        "model.fit(X_train_scaled, y_train)\n",
        "importances = pd.DataFrame(data={\n",
        "    'Attribute': X_train.columns,\n",
        "    'Importance': model.feature_importances_\n",
        "})\n",
        "CNN_importances = statistics.mean(model.feature_importances_)\n",
        "\n",
        "# calculate the importance of GIST feature\n",
        "X_train, X_test, y_train, y_test = train_test_split(GIST_data, train_label, test_size=0.25, random_state=42)\n",
        "\n",
        "ss = MinMaxScaler(feature_range = (0,1))\n",
        "\n",
        "X_train_scaled = ss.fit_transform(X_train)\n",
        "X_test_scaled = ss.transform(X_test)\n",
        "\n",
        "model = XGBClassifier()\n",
        "model.fit(X_train_scaled, y_train)\n",
        "importances = pd.DataFrame(data={\n",
        "    'Attribute': X_train.columns,\n",
        "    'Importance': model.feature_importances_\n",
        "})\n",
        "GIST_importances = statistics.mean(model.feature_importances_)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XLrFmUr-t_GI",
        "outputId": "7cebe593-7df2-491a-83a2-b5150e9211ab"
      },
      "source": [
        "print(CNN_importances)\n",
        "print(GIST_importances)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.00024420023\n",
            "0.001956947\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 338
        },
        "id": "qu6f2cFdYfFl",
        "outputId": "49b83b06-d764-47d5-9c46-18de01096e9f"
      },
      "source": [
        "# test appropriate imputing methods\n",
        "results = list()\n",
        "strategies = ['mean', 'median', 'most_frequent', 'constant']\n",
        "for s in strategies:\n",
        "\t# create the modeling pipeline\n",
        "\tpipeline = Pipeline(steps=[('i', SimpleImputer(strategy=s)), ('m', RandomForestClassifier())])\n",
        "\t# evaluate the model\n",
        "\tcv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)\n",
        "\tscores = cross_val_score(pipeline, pro_data, train_label, scoring='accuracy', cv=cv, n_jobs=-1)\n",
        "\t# store results\n",
        "\tresults.append(scores)\n",
        "\tprint('>%s %.3f (%.3f)' % (s, mean(scores), std(scores)))\n",
        "# plot model performance for comparison\n",
        "plt.boxplot(results, labels=strategies, showmeans=True)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            ">mean 0.751 (0.041)\n",
            ">median 0.749 (0.043)\n",
            ">most_frequent 0.756 (0.043)\n",
            ">constant 0.750 (0.046)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eaMKM5LTaUbm",
        "outputId": "1e753757-d4d4-4eac-e691-902178fd3a81"
      },
      "source": [
        "# get the dataset\n",
        "data_array=data.values\n",
        "scaler = MinMaxScaler(feature_range = (0,1))\n",
        "data_normalization = scaler.fit_transform(data_array)\n",
        "imputer = SimpleImputer(strategy='mean')\n",
        "data_imputing = imputer.fit_transform(data_normalization)\n",
        "\n",
        "# get a list of models to evaluate\n",
        "def get_models():\n",
        "\tmodels = dict()\n",
        "\tfor i in range(1,21):\n",
        "\t\tsteps = [('pca', PCA(n_components=i)), ('m', LogisticRegression())]\n",
        "\t\tmodels[str(i)] = Pipeline(steps=steps)\n",
        "\treturn models\n",
        "\n",
        "# get the models to evaluate\n",
        "models = get_models()\n",
        "# evaluate the models and store results\n",
        "results, names = list(), list()\n",
        "for name, model in models.items():\n",
        "\tscores = evaluate_model(model, data_imputing, train_label)\n",
        "\tresults.append(scores)\n",
        "\tnames.append(name)\n",
        "\tprint('>%s %.3f (%.3f)' % (name, mean(scores), std(scores)))\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            ">1 0.722 (0.038)\n",
            ">2 0.732 (0.042)\n",
            ">3 0.731 (0.042)\n",
            ">4 0.755 (0.044)\n",
            ">5 0.754 (0.043)\n",
            ">6 0.758 (0.044)\n",
            ">7 0.756 (0.047)\n",
            ">8 0.759 (0.047)\n",
            ">9 0.760 (0.046)\n",
            ">10 0.762 (0.045)\n",
            ">11 0.758 (0.049)\n",
            ">12 0.757 (0.048)\n",
            ">13 0.756 (0.045)\n",
            ">14 0.757 (0.050)\n",
            ">15 0.762 (0.047)\n",
            ">16 0.759 (0.042)\n",
            ">17 0.757 (0.044)\n",
            ">18 0.758 (0.046)\n",
            ">19 0.760 (0.048)\n",
            ">20 0.756 (0.047)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vYIP1w09rJ68"
      },
      "source": [
        "# Pre-processing data\n",
        "def dataprocessing(data):\n",
        "\n",
        "  # convert a dataframe into an array\n",
        "  data_array=data.values\n",
        "  # Input normalization: Normalization\n",
        "  scaler = MinMaxScaler(feature_range = (0,1))\n",
        "  data_normalization = scaler.fit_transform(data_array)\n",
        "\n",
        "  # Imputing\n",
        "  imputer = SimpleImputer(strategy='mean')\n",
        "  data_imputing = imputer.fit_transform(data_normalization)\n",
        "  \n",
        "  # dimensionality reduction\n",
        "  pca = PCA(n_components=10)\n",
        "  data_pca = pca.fit_transform(data_imputing) \n",
        "\n",
        "  #draw plot\n",
        "  plt.scatter(data_pca[:,0],data_pca[:,1])\n",
        "  return data_pca\n",
        "\n",
        "\n",
        "# evaluate model\n",
        "def evaluate_model(model, train_x, train_y):\n",
        "  # K fold resampling \n",
        "\tcv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)\n",
        "\tscores = cross_val_score(model, train_x, train_y, scoring='accuracy', cv=cv, n_jobs=-1)\n",
        "\treturn scores\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 264
        },
        "id": "hE7zRLCe7ka-",
        "outputId": "0bb3b142-6b73-45c6-83b4-aa726227b193"
      },
      "source": [
        "# get training data and label data\n",
        "properties = list(data.columns.values)\n",
        "properties.remove('label')\n",
        "pro_data = data[properties]\n",
        "train_label = data['label']\n",
        "train_data=dataprocessing(pro_data)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 826
        },
        "id": "5HO0vTP_bFSB",
        "outputId": "ca0faf50-4338-4f3d-c577-0585ed09c7d2"
      },
      "source": [
        "\n",
        "# test basic classifiers\n",
        "def get_models():\n",
        "  models = dict()\n",
        "  models['Naive Bayes'] =GaussianNB()\n",
        "  models['Logistic Regression'] = LogisticRegression()\n",
        "  models['Linear Discriminant Analysis'] = LinearDiscriminantAnalysis()\n",
        "  models['Random Forest Classifier'] = RandomForestClassifier()\n",
        "  models['Support Vector Machine'] = SVC(probability=True)\n",
        "  models['Multi-layer Perceptron Classifier'] = MLPClassifier()\n",
        "  models['K-Nearest Neighbors'] = KNeighborsClassifier()\n",
        "  return models\n",
        "\n",
        "\n",
        "# get the models to evaluate\n",
        "models = get_models()\n",
        "\n",
        "# evaluate the models and store results\n",
        "results, names = list(), list()\n",
        "for name, model in models.items():\n",
        "  scores = evaluate_model(model, train_data, train_label)\n",
        "  results.append(scores)\n",
        "  names.append(name)\n",
        "  print('Accuracy score of %s : %f (%f)' % (name, mean(scores), std(scores)))\n",
        " \n",
        "# boxplot algorithm comparison\n",
        "# figure related code\n",
        "\n",
        "fig = plt.figure(figsize=(20,10))\n",
        "fig.suptitle('Figure 1. the results of model selection ', fontsize=20)\n",
        "\n",
        "ax = fig.add_subplot(111)\n",
        "ax.boxplot(results, labels=names, showmeans=True)\n",
        "\n",
        "ax.set_title('basic classifiers', fontsize=16, fontweight='bold')\n",
        "ax.set_xlabel('classifiers', fontsize=16, fontweight='bold')\n",
        "ax.set_ylabel('Accuracy score', fontsize=16, fontweight='bold')\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy score of Naive Bayes : 0.743135 (0.042854)\n",
            "Accuracy score of Logistic Regression : 0.753289 (0.048485)\n",
            "Accuracy score of Linear Discriminant Analysis : 0.752862 (0.047738)\n",
            "Accuracy score of Random Forest Classifier : 0.736790 (0.041989)\n",
            "Accuracy score of Support Vector Machine : 0.749903 (0.048099)\n",
            "Accuracy score of Multi-layer Perceptron Classifier : 0.716959 (0.049724)\n",
            "Accuracy score of K-Nearest Neighbors : 0.697517 (0.041633)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'Accuracy score')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1ot6c-nR46dQ",
        "outputId": "01bf6283-f959-4b87-97dc-3c4022bfe895"
      },
      "source": [
        "# feature selection with RFE\n",
        "def get_models():\n",
        "  models = dict()\n",
        "  rfe = RFE(estimator=GaussianNB(), n_features_to_select=3)\n",
        "  model = LogisticRegression()\n",
        "  models['Naive Bayes'] = Pipeline(steps=[('s',rfe),('m',model)])\n",
        "  # \n",
        "  rfe = RFE(estimator=LogisticRegression(), n_features_to_select=3)\n",
        "  model = LogisticRegression()\n",
        "  models['Logistic Regression'] = Pipeline(steps=[('s',rfe),('m',model)])\n",
        "  # \n",
        "  rfe = RFE(estimator=LogisticRegression(), n_features_to_select=3)\n",
        "  model = estimator=SVC(probability=True)\n",
        "  models['Support Vector Machine'] = Pipeline(steps=[('s',rfe),('m',model)])\n",
        "\t# \n",
        "  rfe = RFE(estimator=LogisticRegression(), n_features_to_select=3)\n",
        "  model = RandomForestClassifier()\n",
        "  models['Random Forest Classifier'] = Pipeline(steps=[('s',rfe),('m',model)])\n",
        "\t# \n",
        "  rfe = RFE(estimator= LogisticRegression(), n_features_to_select=3)\n",
        "  model = MLPClassifier()\n",
        "  models[' Multi-layer Perceptron Classifier'] = Pipeline(steps=[('s',rfe),('m',model)])\n",
        "  return models\n",
        "\n",
        "\n",
        "\n",
        "# get the models to evaluate\n",
        "models = get_models()\n",
        "\n",
        "# evaluate the models and store results\n",
        "results, names = list(), list()\n",
        "for name, model in models.items():\n",
        "  scores = evaluate_model(model, train_data, train_label)\n",
        "  results.append(scores)\n",
        "  names.append(name)\n",
        "  print('Accuracy score of %s : %f (%f)' % (name, mean(scores), std(scores)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy score of Naive Bayes : nan (nan)\n",
            "Accuracy score of Logistic Regression : 0.742719 (0.037677)\n",
            "Accuracy score of Support Vector Machine : 0.740587 (0.042669)\n",
            "Accuracy score of Random Forest Classifier : 0.714822 (0.052353)\n",
            "Accuracy score of  Multi-layer Perceptron Classifier : 0.741870 (0.042548)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "46oy6N8f4v7I",
        "outputId": "1e856b45-1fb3-48eb-a757-06b43b6224b9"
      },
      "source": [
        "model = LogisticRegression()\n",
        "\n",
        "param_grid = [\n",
        "    {\n",
        "    'penalty' : ['l1', 'l2'],\n",
        "    'C' : np.logspace(-4, 4, 20),\n",
        "    'solver' : ['newton-cg', 'lbfgs', 'liblinear'],\n",
        "    'max_iter' : [100,110,120,130,140],\n",
        "    'dual':[True,False]\n",
        "    }\n",
        "]\n",
        "\n",
        "# define grid search\n",
        "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1)\n",
        "grid_search  = GridSearchCV(estimator=model, param_grid = param_grid, n_jobs=-1, cv=cv, scoring='accuracy')\n",
        "grid_result = grid_search.fit(train_data, train_label)\n",
        "\n",
        "# summarize results\n",
        "print(\"Best Logistic Regression: %f using %s\" % (grid_result.best_score_, grid_result.best_params_))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Best Logistic Regression: 0.754549 using {'C': 0.08858667904100823, 'dual': False, 'max_iter': 100, 'penalty': 'l2', 'solver': 'newton-cg'}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YGYTcOZfK5V4",
        "outputId": "ac9ec9d4-056c-4c7a-96fe-9c23292a230b"
      },
      "source": [
        "nb_classifier = GaussianNB()\n",
        "\n",
        "params_NB = {'var_smoothing': np.logspace(0,-9, num=100)}\n",
        "gs_NB = GridSearchCV(estimator=nb_classifier, param_grid=params_NB,cv=cv,verbose=1,scoring='accuracy') \n",
        "gs_NB.fit(train_data, train_label)\n",
        "\n",
        "gs_NB.best_params_"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fitting 30 folds for each of 100 candidates, totalling 3000 fits\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
            "[Parallel(n_jobs=1)]: Done 3000 out of 3000 | elapsed:    5.1s finished\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'var_smoothing': 0.1}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zjVyh_6iFIyv",
        "outputId": "12b8f199-c9e0-4d6e-8219-520fa7b9de65"
      },
      "source": [
        "# define models and parameters\n",
        "model = MLPClassifier(max_iter=100)\n",
        "param_grid = [\n",
        "    { 'hidden_layer_sizes': [(50,50,50), (50,100,50), (100,)],\n",
        "    'activation': ['tanh', 'relu'],\n",
        "    'solver': ['sgd', 'adam'],\n",
        "    'alpha': [0.0001, 0.05],\n",
        "    'learning_rate': ['constant','adaptive'],\n",
        "     }\n",
        "]\n",
        "\n",
        "# define grid search\n",
        "grid_search  = GridSearchCV(estimator=model, param_grid = param_grid, n_jobs=-1, cv=5, scoring='accuracy')\n",
        "grid_result = grid_search.fit(train_data, train_label)\n",
        "\n",
        "# summarize results\n",
        "print(\"Best MLPClassifier: %f using %s\" % (grid_result.best_score_, grid_result.best_params_))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Best MLPClassifier: 0.759203 using {'activation': 'tanh', 'alpha': 0.05, 'hidden_layer_sizes': (50, 50, 50), 'learning_rate': 'constant', 'solver': 'sgd'}\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (100) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0jk0DYDCwvl1"
      },
      "source": [
        "def get_models():\n",
        "  models = dict()\n",
        "  models['Naive Bayes2'] =GaussianNB(var_smoothing=0.08111308307896872)\n",
        "  models['Logistic Regression2'] = LogisticRegression(penalty='l2',C=0.012742749857031334,solver='liblinear',max_iter = 100)\n",
        "  models['Linear Discriminant Analysis'] = LinearDiscriminantAnalysis()\n",
        "  models['Multi-layer Perceptron Classifier2'] = MLPClassifier(max_iter=100,activation='tanh',alpha= 0.0001,hidden_layer_sizes= (100,),learning_rate='adaptive', solver='adam')\n",
        "  return models"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 772
        },
        "id": "zM69q7-V9ajP",
        "outputId": "23185b2f-e2e2-4235-bbf9-be0c13eac576"
      },
      "source": [
        "\n",
        "# get the models to evaluate\n",
        "models = get_models()\n",
        "\n",
        "# evaluate the models and store results\n",
        "results, names = list(), list()\n",
        "for name, model in models.items():\n",
        "  scores = evaluate_model(model, train_data, train_label)\n",
        "  results.append(scores)\n",
        "  names.append(name)\n",
        "  print('Accuracy score of %s : %f (%f)' % (name, mean(scores), std(scores)))\n",
        " \n",
        "# boxplot algorithm comparison\n",
        "# figure related code\n",
        "\n",
        "fig = plt.figure(figsize=(20,10))\n",
        "fig.suptitle('Figure 2. the results of model selection ', fontsize=20)\n",
        "\n",
        "ax = fig.add_subplot(111)\n",
        "ax.boxplot(results, labels=names, showmeans=True)\n",
        "\n",
        "ax.set_title('basic classifiers', fontsize=16, fontweight='bold')\n",
        "ax.set_xlabel('classifiers', fontsize=16, fontweight='bold')\n",
        "ax.set_ylabel('Accuracy score', fontsize=16, fontweight='bold')\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy score of Naive Bayes2 : 0.743135 (0.041072)\n",
            "Accuracy score of Logistic Regression2 : 0.751163 (0.048969)\n",
            "Accuracy score of Linear Discriminant Analysis : 0.752862 (0.047738)\n",
            "Accuracy score of Multi-layer Perceptron Classifier2 : 0.753267 (0.045650)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'Accuracy score')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 27
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 280
        },
        "id": "ueakHi6J5XGH",
        "outputId": "f8b9b892-e702-4450-c6f2-e7e66ae1c6f7"
      },
      "source": [
        "# test basic classifiers\n",
        "def get_models():\n",
        "  models = dict()\n",
        "  models['Logistic Regression'] = LogisticRegression()\n",
        "  models['Logistic Regression2'] = LogisticRegression(penalty='l2')\n",
        "  models['Logistic Regression3'] = LogisticRegression(penalty='l2',C=0.012742749857031334)\n",
        "  models['Logistic Regression4'] = LogisticRegression(penalty='l2',C=0.012742749857031334,solver='newton-cg')\n",
        "  models['Logistic Regression5'] = LogisticRegression(penalty='l1',C=0.012742749857031334,solver='liblinear')\n",
        "  models['Logistic Regression6'] = LogisticRegression(penalty='l2',C=0.012742749857031334,solver='liblinear')\n",
        "  return models\n",
        "\n",
        "\n",
        "# get the models to evaluate\n",
        "models = get_models()\n",
        "\n",
        "for name, model in models.items():\n",
        "    (train_sizes,\n",
        "     train_scores,\n",
        "     test_scores) = learning_curve(model, train_data, train_label, cv=5)\n",
        "    test_mean = np.mean(test_scores, axis=1)\n",
        "    plt.plot(train_sizes, test_mean, label='learning curve of %s ' % (name))\n",
        "\n",
        "plt.title(\"Figure 3. learning curve for differnet Logistic Regression Classifiers\")\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "D_VmziK8U99Q",
        "outputId": "da1d9db2-eab9-4f2e-e08f-73ce59f2c6ea"
      },
      "source": [
        "# draw learning curve for logistic regression\n",
        "model=LogisticRegression(penalty='l2',C=0.012742749857031334,solver='liblinear',max_iter = 100)\n",
        "\n",
        "# cv is the number of folds while performing Cross Validation\n",
        "sizes, training_scores, testing_scores = learning_curve(model, train_data, train_label, cv=10, scoring='accuracy', train_sizes=np.linspace(0.01, 1.0, 50))\n",
        "  \n",
        "# Mean and Standard Deviation of training scores\n",
        "mean_training = np.mean(training_scores, axis=1)\n",
        "Standard_Deviation_training = np.std(training_scores, axis=1)\n",
        "  \n",
        "# Mean and Standard Deviation of testing scores\n",
        "mean_testing = np.mean(testing_scores, axis=1)\n",
        "Standard_Deviation_testing = np.std(testing_scores, axis=1)\n",
        "  \n",
        "# dotted blue line is for training scores and green line is for cross-validation score\n",
        "plt.plot(sizes, mean_training, '--', color=\"b\",  label=\"Training score\")\n",
        "plt.plot(sizes, mean_testing, color=\"g\", label=\"Cross-validation score\")\n",
        "  \n",
        "# Drawing plot\n",
        "plt.title(\"Figure 4. learning curve for final Logistic Regression Classifier\")\n",
        "plt.xlabel(\"Training Set Size\"), plt.ylabel(\"Accuracy Score\"), plt.legend(loc=\"best\")\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E8ogyWEX5Ei-"
      },
      "source": [
        "test=pd.read_csv('/content/drive/My Drive/FML_Assignment/test.csv')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 439
        },
        "id": "rDZuXFbh5OZf",
        "outputId": "3a7282bc-c898-41fc-efc4-07b454089901"
      },
      "source": [
        "test"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>CNNs</th>\n",
              "      <th>CNNs.1</th>\n",
              "      <th>CNNs.2</th>\n",
              "      <th>CNNs.3</th>\n",
              "      <th>CNNs.4</th>\n",
              "      <th>CNNs.5</th>\n",
              "      <th>CNNs.6</th>\n",
              "      <th>CNNs.7</th>\n",
              "      <th>CNNs.8</th>\n",
              "      <th>CNNs.9</th>\n",
              "      <th>CNNs.10</th>\n",
              "      <th>CNNs.11</th>\n",
              "      <th>CNNs.12</th>\n",
              "      <th>CNNs.13</th>\n",
              "      <th>CNNs.14</th>\n",
              "      <th>CNNs.15</th>\n",
              "      <th>CNNs.16</th>\n",
              "      <th>CNNs.17</th>\n",
              "      <th>CNNs.18</th>\n",
              "      <th>CNNs.19</th>\n",
              "      <th>CNNs.20</th>\n",
              "      <th>CNNs.21</th>\n",
              "      <th>CNNs.22</th>\n",
              "      <th>CNNs.23</th>\n",
              "      <th>CNNs.24</th>\n",
              "      <th>CNNs.25</th>\n",
              "      <th>CNNs.26</th>\n",
              "      <th>CNNs.27</th>\n",
              "      <th>CNNs.28</th>\n",
              "      <th>CNNs.29</th>\n",
              "      <th>CNNs.30</th>\n",
              "      <th>CNNs.31</th>\n",
              "      <th>CNNs.32</th>\n",
              "      <th>CNNs.33</th>\n",
              "      <th>CNNs.34</th>\n",
              "      <th>CNNs.35</th>\n",
              "      <th>CNNs.36</th>\n",
              "      <th>CNNs.37</th>\n",
              "      <th>CNNs.38</th>\n",
              "      <th>CNNs.39</th>\n",
              "      <th>...</th>\n",
              "      <th>GIST.472</th>\n",
              "      <th>GIST.473</th>\n",
              "      <th>GIST.474</th>\n",
              "      <th>GIST.475</th>\n",
              "      <th>GIST.476</th>\n",
              "      <th>GIST.477</th>\n",
              "      <th>GIST.478</th>\n",
              "      <th>GIST.479</th>\n",
              "      <th>GIST.480</th>\n",
              "      <th>GIST.481</th>\n",
              "      <th>GIST.482</th>\n",
              "      <th>GIST.483</th>\n",
              "      <th>GIST.484</th>\n",
              "      <th>GIST.485</th>\n",
              "      <th>GIST.486</th>\n",
              "      <th>GIST.487</th>\n",
              "      <th>GIST.488</th>\n",
              "      <th>GIST.489</th>\n",
              "      <th>GIST.490</th>\n",
              "      <th>GIST.491</th>\n",
              "      <th>GIST.492</th>\n",
              "      <th>GIST.493</th>\n",
              "      <th>GIST.494</th>\n",
              "      <th>GIST.495</th>\n",
              "      <th>GIST.496</th>\n",
              "      <th>GIST.497</th>\n",
              "      <th>GIST.498</th>\n",
              "      <th>GIST.499</th>\n",
              "      <th>GIST.500</th>\n",
              "      <th>GIST.501</th>\n",
              "      <th>GIST.502</th>\n",
              "      <th>GIST.503</th>\n",
              "      <th>GIST.504</th>\n",
              "      <th>GIST.505</th>\n",
              "      <th>GIST.506</th>\n",
              "      <th>GIST.507</th>\n",
              "      <th>GIST.508</th>\n",
              "      <th>GIST.509</th>\n",
              "      <th>GIST.510</th>\n",
              "      <th>GIST.511</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.641460</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.010263</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.09380</td>\n",
              "      <td>0.67192</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.146000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.23594</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.023848</td>\n",
              "      <td>0.410790</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.004334</td>\n",
              "      <td>0.006872</td>\n",
              "      <td>0.007383</td>\n",
              "      <td>0.023129</td>\n",
              "      <td>0.015715</td>\n",
              "      <td>0.014801</td>\n",
              "      <td>0.005751</td>\n",
              "      <td>0.018266</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.027327</td>\n",
              "      <td>0.015378</td>\n",
              "      <td>0.018204</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.006307</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.050680</td>\n",
              "      <td>0.003120</td>\n",
              "      <td>0.005427</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.083366</td>\n",
              "      <td>0.013520</td>\n",
              "      <td>0.014224</td>\n",
              "      <td>0.013778</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.024670</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.020640</td>\n",
              "      <td>0.019328</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.048529</td>\n",
              "      <td>0.013987</td>\n",
              "      <td>0.009278</td>\n",
              "      <td>0.011328</td>\n",
              "      <td>0.010538</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.013829</td>\n",
              "      <td>0.052630</td>\n",
              "      <td>0.035465</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.058088</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.78386</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.008242</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.783720</td>\n",
              "      <td>1.26100</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.83732</td>\n",
              "      <td>0.28559</td>\n",
              "      <td>0.37439</td>\n",
              "      <td>0.34635</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.59925</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.63904</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.004064</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.017208</td>\n",
              "      <td>0.024160</td>\n",
              "      <td>0.043782</td>\n",
              "      <td>0.029639</td>\n",
              "      <td>0.037706</td>\n",
              "      <td>0.025978</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.045894</td>\n",
              "      <td>0.027285</td>\n",
              "      <td>0.012793</td>\n",
              "      <td>0.021836</td>\n",
              "      <td>0.024630</td>\n",
              "      <td>0.007313</td>\n",
              "      <td>0.011662</td>\n",
              "      <td>0.012753</td>\n",
              "      <td>0.030091</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.017599</td>\n",
              "      <td>0.049867</td>\n",
              "      <td>0.011505</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.020126</td>\n",
              "      <td>0.015193</td>\n",
              "      <td>0.019641</td>\n",
              "      <td>0.112840</td>\n",
              "      <td>0.147830</td>\n",
              "      <td>0.033717</td>\n",
              "      <td>0.020202</td>\n",
              "      <td>0.022431</td>\n",
              "      <td>0.047402</td>\n",
              "      <td>0.015783</td>\n",
              "      <td>0.008880</td>\n",
              "      <td>0.026375</td>\n",
              "      <td>0.044499</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.032760</td>\n",
              "      <td>0.019160</td>\n",
              "      <td>0.042316</td>\n",
              "      <td>0.014976</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.38465</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.70955</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.80505</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>1.14640</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.73152</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>...</td>\n",
              "      <td>0.026901</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.029906</td>\n",
              "      <td>0.013477</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.026310</td>\n",
              "      <td>0.024671</td>\n",
              "      <td>0.051806</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.018258</td>\n",
              "      <td>0.028155</td>\n",
              "      <td>0.015256</td>\n",
              "      <td>0.028930</td>\n",
              "      <td>0.020556</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.035673</td>\n",
              "      <td>0.018933</td>\n",
              "      <td>0.054751</td>\n",
              "      <td>0.029442</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.025339</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.033817</td>\n",
              "      <td>0.033160</td>\n",
              "      <td>0.020813</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.008986</td>\n",
              "      <td>0.017565</td>\n",
              "      <td>0.026905</td>\n",
              "      <td>0.014389</td>\n",
              "      <td>0.022969</td>\n",
              "      <td>0.027598</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.024075</td>\n",
              "      <td>0.031379</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.71303</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.571850</td>\n",
              "      <td>0.20295</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.18072</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.323420</td>\n",
              "      <td>0.028327</td>\n",
              "      <td>0.64687</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.03384</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.045159</td>\n",
              "      <td>0.13076</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.190680</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.50899</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.24303</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.016763</td>\n",
              "      <td>0.030263</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.020161</td>\n",
              "      <td>0.003834</td>\n",
              "      <td>0.011930</td>\n",
              "      <td>0.005787</td>\n",
              "      <td>0.013193</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.018324</td>\n",
              "      <td>0.001617</td>\n",
              "      <td>0.039116</td>\n",
              "      <td>0.024671</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.017718</td>\n",
              "      <td>0.001878</td>\n",
              "      <td>0.013104</td>\n",
              "      <td>0.010996</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.011977</td>\n",
              "      <td>0.018151</td>\n",
              "      <td>0.024082</td>\n",
              "      <td>0.026165</td>\n",
              "      <td>0.020368</td>\n",
              "      <td>0.031549</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.026506</td>\n",
              "      <td>0.005279</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.028113</td>\n",
              "      <td>0.035376</td>\n",
              "      <td>0.001820</td>\n",
              "      <td>0.003180</td>\n",
              "      <td>0.016161</td>\n",
              "      <td>0.000948</td>\n",
              "      <td>0.022530</td>\n",
              "      <td>0.013404</td>\n",
              "      <td>0.022238</td>\n",
              "      <td>0.039642</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.50370</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.47737</td>\n",
              "      <td>0.43069</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.24137</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.5146</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.87694</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.81024</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.76022</td>\n",
              "      <td>1.27470</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.410130</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.016372</td>\n",
              "      <td>0.013736</td>\n",
              "      <td>0.023753</td>\n",
              "      <td>0.036873</td>\n",
              "      <td>0.078999</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.014637</td>\n",
              "      <td>0.026880</td>\n",
              "      <td>0.013466</td>\n",
              "      <td>0.084849</td>\n",
              "      <td>0.055964</td>\n",
              "      <td>0.007813</td>\n",
              "      <td>0.028473</td>\n",
              "      <td>0.007077</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.024678</td>\n",
              "      <td>0.032560</td>\n",
              "      <td>0.013846</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.054825</td>\n",
              "      <td>0.035741</td>\n",
              "      <td>0.029675</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.042840</td>\n",
              "      <td>0.024686</td>\n",
              "      <td>0.030309</td>\n",
              "      <td>0.043754</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.013753</td>\n",
              "      <td>0.016204</td>\n",
              "      <td>0.006194</td>\n",
              "      <td>0.007046</td>\n",
              "      <td>0.027577</td>\n",
              "      <td>0.023283</td>\n",
              "      <td>0.011591</td>\n",
              "      <td>0.086180</td>\n",
              "      <td>0.023861</td>\n",
              "      <td>0.040862</td>\n",
              "      <td>0.021848</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2162</th>\n",
              "      <td>0.34403</td>\n",
              "      <td>1.07600</td>\n",
              "      <td>0.38587</td>\n",
              "      <td>0.436380</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.29000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.412180</td>\n",
              "      <td>0.31762</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.29000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.499200</td>\n",
              "      <td>0.77590</td>\n",
              "      <td>1.28690</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.52901</td>\n",
              "      <td>0.082485</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.56875</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.28816</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.56325</td>\n",
              "      <td>0.84117</td>\n",
              "      <td>0.59226</td>\n",
              "      <td>1.641200</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.021072</td>\n",
              "      <td>0.783620</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.011784</td>\n",
              "      <td>0.017623</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.007162</td>\n",
              "      <td>0.006442</td>\n",
              "      <td>0.005603</td>\n",
              "      <td>0.005451</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.014814</td>\n",
              "      <td>0.004312</td>\n",
              "      <td>0.010977</td>\n",
              "      <td>0.007085</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.007252</td>\n",
              "      <td>0.011463</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.035779</td>\n",
              "      <td>0.063967</td>\n",
              "      <td>0.017778</td>\n",
              "      <td>0.007211</td>\n",
              "      <td>0.085765</td>\n",
              "      <td>0.021080</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.021570</td>\n",
              "      <td>0.035117</td>\n",
              "      <td>0.016950</td>\n",
              "      <td>0.004441</td>\n",
              "      <td>0.003119</td>\n",
              "      <td>0.020952</td>\n",
              "      <td>0.012463</td>\n",
              "      <td>0.001605</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.012048</td>\n",
              "      <td>0.008157</td>\n",
              "      <td>0.024373</td>\n",
              "      <td>0.005569</td>\n",
              "      <td>0.023510</td>\n",
              "      <td>0.021689</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2163</th>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.20164</td>\n",
              "      <td>0.35045</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.60092</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.15429</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.41600</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.781840</td>\n",
              "      <td>0.98618</td>\n",
              "      <td>1.67450</td>\n",
              "      <td>0.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.10377</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.12200</td>\n",
              "      <td>1.26360</td>\n",
              "      <td>1.053300</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.163700</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.001263</td>\n",
              "      <td>0.008285</td>\n",
              "      <td>0.017906</td>\n",
              "      <td>0.004717</td>\n",
              "      <td>0.001204</td>\n",
              "      <td>0.001938</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.001584</td>\n",
              "      <td>0.005312</td>\n",
              "      <td>0.000659</td>\n",
              "      <td>0.000994</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.006470</td>\n",
              "      <td>0.002293</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.012430</td>\n",
              "      <td>0.001368</td>\n",
              "      <td>0.001240</td>\n",
              "      <td>0.005526</td>\n",
              "      <td>0.000524</td>\n",
              "      <td>0.001339</td>\n",
              "      <td>0.029816</td>\n",
              "      <td>0.001284</td>\n",
              "      <td>0.012110</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.027986</td>\n",
              "      <td>0.039157</td>\n",
              "      <td>0.001328</td>\n",
              "      <td>0.001296</td>\n",
              "      <td>0.002192</td>\n",
              "      <td>0.011031</td>\n",
              "      <td>0.010472</td>\n",
              "      <td>0.002153</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.009954</td>\n",
              "      <td>0.001591</td>\n",
              "      <td>0.004829</td>\n",
              "      <td>0.006363</td>\n",
              "      <td>0.006330</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2164</th>\n",
              "      <td>0.49384</td>\n",
              "      <td>1.31680</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.159750</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>1.259400</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>1.5129</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.068000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.22156</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.91693</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.031945</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.015190</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.056224</td>\n",
              "      <td>0.023376</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.011274</td>\n",
              "      <td>0.012663</td>\n",
              "      <td>0.038156</td>\n",
              "      <td>0.028165</td>\n",
              "      <td>0.008315</td>\n",
              "      <td>0.018845</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.007956</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.055204</td>\n",
              "      <td>0.014900</td>\n",
              "      <td>0.017451</td>\n",
              "      <td>0.021934</td>\n",
              "      <td>0.048078</td>\n",
              "      <td>0.036569</td>\n",
              "      <td>0.037189</td>\n",
              "      <td>0.021401</td>\n",
              "      <td>0.038172</td>\n",
              "      <td>0.033874</td>\n",
              "      <td>0.056149</td>\n",
              "      <td>0.082901</td>\n",
              "      <td>0.020904</td>\n",
              "      <td>0.017688</td>\n",
              "      <td>0.016386</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.020838</td>\n",
              "      <td>0.006681</td>\n",
              "      <td>0.017014</td>\n",
              "      <td>0.040438</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.031057</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.023415</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2165</th>\n",
              "      <td>0.13996</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.80403</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.068594</td>\n",
              "      <td>0.302900</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.52423</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.034848</td>\n",
              "      <td>0.63780</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.31555</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.018602</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.71044</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.683100</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.073706</td>\n",
              "      <td>0.471600</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.017545</td>\n",
              "      <td>0.029115</td>\n",
              "      <td>0.021795</td>\n",
              "      <td>0.028140</td>\n",
              "      <td>0.017098</td>\n",
              "      <td>0.021135</td>\n",
              "      <td>0.052297</td>\n",
              "      <td>0.059540</td>\n",
              "      <td>0.028169</td>\n",
              "      <td>0.039859</td>\n",
              "      <td>0.023326</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.048557</td>\n",
              "      <td>0.016664</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.014002</td>\n",
              "      <td>0.045326</td>\n",
              "      <td>0.101300</td>\n",
              "      <td>0.031914</td>\n",
              "      <td>0.023953</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.024437</td>\n",
              "      <td>0.017005</td>\n",
              "      <td>0.022638</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.043686</td>\n",
              "      <td>0.015637</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.026577</td>\n",
              "      <td>0.022878</td>\n",
              "      <td>0.021351</td>\n",
              "      <td>0.020791</td>\n",
              "      <td>0.038496</td>\n",
              "      <td>0.023499</td>\n",
              "      <td>0.050489</td>\n",
              "      <td>0.031861</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2166</th>\n",
              "      <td>NaN</td>\n",
              "      <td>0.24467</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.040481</td>\n",
              "      <td>0.39675</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.036280</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.248990</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.38767</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.788700</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.004841</td>\n",
              "      <td>0.016311</td>\n",
              "      <td>0.013265</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.012056</td>\n",
              "      <td>0.035664</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.044097</td>\n",
              "      <td>0.016025</td>\n",
              "      <td>0.002979</td>\n",
              "      <td>0.017299</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.009073</td>\n",
              "      <td>0.008246</td>\n",
              "      <td>0.004716</td>\n",
              "      <td>0.053546</td>\n",
              "      <td>0.009458</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.034158</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.042834</td>\n",
              "      <td>0.017164</td>\n",
              "      <td>0.043633</td>\n",
              "      <td>0.019243</td>\n",
              "      <td>0.076615</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.021775</td>\n",
              "      <td>0.019624</td>\n",
              "      <td>0.007226</td>\n",
              "      <td>0.041641</td>\n",
              "      <td>0.013943</td>\n",
              "      <td>0.007421</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.043014</td>\n",
              "      <td>0.007863</td>\n",
              "      <td>0.029368</td>\n",
              "      <td>0.023330</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.003198</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>2167 rows × 4608 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "         CNNs   CNNs.1   CNNs.2  ...  GIST.509  GIST.510  GIST.511\n",
              "0     0.00000  0.00000  0.00000  ...  0.035465       NaN  0.058088\n",
              "1     0.00000  0.78386  0.00000  ...  0.019160  0.042316  0.014976\n",
              "2     0.38465  0.00000  0.00000  ...       NaN  0.024075  0.031379\n",
              "3     0.00000  0.71303  0.00000  ...  0.013404  0.022238  0.039642\n",
              "4         NaN  0.50370      NaN  ...  0.023861  0.040862  0.021848\n",
              "...       ...      ...      ...  ...       ...       ...       ...\n",
              "2162  0.34403  1.07600  0.38587  ...  0.005569  0.023510  0.021689\n",
              "2163  0.00000  0.20164  0.35045  ...  0.006363  0.006330       NaN\n",
              "2164  0.49384  1.31680      NaN  ...       NaN  0.023415       NaN\n",
              "2165  0.13996  0.00000      NaN  ...  0.023499  0.050489  0.031861\n",
              "2166      NaN  0.24467  0.00000  ...  0.023330       NaN  0.003198\n",
              "\n",
              "[2167 rows x 4608 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "QHYhNN9N7GdZ",
        "outputId": "76bc7141-2f73-472f-b8d9-3f88329e801b"
      },
      "source": [
        "test_data=dataprocessing(test)\n",
        "# fit final model\n",
        "model=LogisticRegression(penalty='l2',C=0.012742749857031334,solver='liblinear',max_iter = 100)\n",
        "model.fit(train_data,train_label)\n",
        "predictions = model.predict(test_data)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 17
        },
        "id": "QfOhJjFP90Vb",
        "outputId": "d118e6d8-4c5c-4236-ad61-7dbef8ee62ba"
      },
      "source": [
        "predictions\n",
        "prediction = pd.DataFrame(predictions, columns=['predictions']).to_csv('prediction.csv')\n",
        "from google.colab import files\n",
        "files.download(\"prediction.csv\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "    async function download(id, filename, size) {\n",
              "      if (!google.colab.kernel.accessAllowed) {\n",
              "        return;\n",
              "      }\n",
              "      const div = document.createElement('div');\n",
              "      const label = document.createElement('label');\n",
              "      label.textContent = `Downloading \"${filename}\": `;\n",
              "      div.appendChild(label);\n",
              "      const progress = document.createElement('progress');\n",
              "      progress.max = size;\n",
              "      div.appendChild(progress);\n",
              "      document.body.appendChild(div);\n",
              "\n",
              "      const buffers = [];\n",
              "      let downloaded = 0;\n",
              "\n",
              "      const channel = await google.colab.kernel.comms.open(id);\n",
              "      // Send a message to notify the kernel that we're ready.\n",
              "      channel.send({})\n",
              "\n",
              "      for await (const message of channel.messages) {\n",
              "        // Send a message to notify the kernel that we're ready.\n",
              "        channel.send({})\n",
              "        if (message.buffers) {\n",
              "          for (const buffer of message.buffers) {\n",
              "            buffers.push(buffer);\n",
              "            downloaded += buffer.byteLength;\n",
              "            progress.value = downloaded;\n",
              "          }\n",
              "        }\n",
              "      }\n",
              "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
              "      const a = document.createElement('a');\n",
              "      a.href = window.URL.createObjectURL(blob);\n",
              "      a.download = filename;\n",
              "      div.appendChild(a);\n",
              "      a.click();\n",
              "      div.remove();\n",
              "    }\n",
              "  "
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "download(\"download_b6a9cb26-70dc-4e6e-af9a-487060d308ce\", \"prediction.csv\", 18406)"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JdkOlopUDVxD"
      },
      "source": [
        "test"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "EI_A04I15wUV",
        "outputId": "e2ed1226-c36c-4a44-afb4-58e80661c3c8"
      },
      "source": [
        "from sklearn.ensemble import AdaBoostClassifier\n",
        "model = AdaBoostClassifier()\n",
        "param_grid = {'n_estimators': [10, 50, 100, 500],\n",
        " 'learning_rate': [0.0001, 0.001, 0.01, 0.1, 1.0]\n",
        " }\n",
        "\n",
        "# define grid search\n",
        "grid_search  = GridSearchCV(estimator=model, param_grid = param_grid, n_jobs=-1, cv=5, scoring='accuracy')\n",
        "grid_result = grid_search.fit(train_data, train_label)\n",
        "\n",
        "# summarize results\n",
        "print(\"Best AdaBoostClassifier: %f using %s\" % (grid_result.best_score_, grid_result.best_params_))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Best AdaBoostClassifier: 0.749069 using {'learning_rate': 0.1, 'n_estimators': 100}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "Qb4QUlrZ7VxF",
        "outputId": "3c3caa3d-741d-49e9-be45-751920905464"
      },
      "source": [
        "from sklearn.model_selection import learning_curve\n",
        "  \n",
        "modelmodel = AdaBoostClassifier(learning_rate=0.1,n_estimators= 100)\n",
        "# cv is the number of folds while performing Cross Validation\n",
        "sizes, training_scores, testing_scores = learning_curve(model, train_data, train_label, cv=10, scoring='accuracy', train_sizes=np.linspace(0.01, 1.0, 50))\n",
        "  \n",
        "# Mean and Standard Deviation of training scores\n",
        "mean_training = np.mean(training_scores, axis=1)\n",
        "Standard_Deviation_training = np.std(training_scores, axis=1)\n",
        "  \n",
        "# Mean and Standard Deviation of testing scores\n",
        "mean_testing = np.mean(testing_scores, axis=1)\n",
        "Standard_Deviation_testing = np.std(testing_scores, axis=1)\n",
        "  \n",
        "# dotted blue line is for training scores and green line is for cross-validation score\n",
        "plt.plot(sizes, mean_training, '--', color=\"b\",  label=\"Training score\")\n",
        "plt.plot(sizes, mean_testing, color=\"g\", label=\"Cross-validation score\")\n",
        "  \n",
        "# Drawing plot\n",
        "plt.title(\"Figure 5. learning curve for AdaBoost Classifier Classifier\")\n",
        "plt.xlabel(\"Training Set Size\"), plt.ylabel(\"Accuracy Score\"), plt.legend(loc=\"best\")\n",
        "plt.show()"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "SahMtroWWmv6",
        "outputId": "d9970754-7fc1-424d-b652-92b926ae49b0"
      },
      "source": [
        "# test data include low confidence\n",
        "data2=training2\n",
        "# get training data and label data\n",
        "properties2 = list(data.columns.values)\n",
        "properties2.remove('label')\n",
        "pro_data2 = data2[properties]\n",
        "train_label2 = data2['label']\n",
        "train_data2=dataprocessing(pro_data2)\n",
        "\n",
        "# get the models to evaluate\n",
        "model = LogisticRegression(penalty='l2',C=0.012742749857031334,solver='liblinear',max_iter = 100)\n",
        "# evaluate the models and store results\n",
        "scores = evaluate_model(model, train_data, train_label)\n",
        "scores2 = evaluate_model(model, train_data2, train_label2)\n",
        "print('accuracy score of test data without low confidence: %f ' % (mean(scores)))\n",
        "print('accuracy score of test data include low confidence : %f ' % (mean(scores2)))"
      ],
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "accuracy score of test data without low confidence: 0.751163 \n",
            "accuracy score of test data include low confidence : 0.750741 \n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}