Schlumberger_solution.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Schlumberger.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "# устанка catboost\n",
        "pip install catboost"
      ],
      "metadata": {
        "id": "Uu2LfZFiVstm",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e82bde46-f240-4944-82b2-df2f25ac6763"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting catboost\n",
            "  Downloading catboost-1.0.4-cp37-none-manylinux1_x86_64.whl (76.1 MB)\n",
            "\u001b[K     |████████████████████████████████| 76.1 MB 1.3 MB/s \n",
            "\u001b[?25hRequirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from catboost) (1.4.1)\n",
            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from catboost) (3.2.2)\n",
            "Requirement already satisfied: graphviz in /usr/local/lib/python3.7/dist-packages (from catboost) (0.10.1)\n",
            "Requirement already satisfied: numpy>=1.16.0 in /usr/local/lib/python3.7/dist-packages (from catboost) (1.21.5)\n",
            "Requirement already satisfied: plotly in /usr/local/lib/python3.7/dist-packages (from catboost) (5.5.0)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.7/dist-packages (from catboost) (1.15.0)\n",
            "Requirement already satisfied: pandas>=0.24.0 in /usr/local/lib/python3.7/dist-packages (from catboost) (1.3.5)\n",
            "Requirement already satisfied: python-dateutil>=2.7.3 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.24.0->catboost) (2.8.2)\n",
            "Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.24.0->catboost) (2018.9)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->catboost) (0.11.0)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->catboost) (1.3.2)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->catboost) (3.0.7)\n",
            "Requirement already satisfied: tenacity>=6.2.0 in /usr/local/lib/python3.7/dist-packages (from plotly->catboost) (8.0.1)\n",
            "Installing collected packages: catboost\n",
            "Successfully installed catboost-1.0.4\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 158,
      "metadata": {
        "id": "xUe-1KhWeY8J"
      },
      "outputs": [],
      "source": [
        "# импортируем нужные библиотеки\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "from scipy import stats\n",
        "from sklearn.impute import KNNImputer\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import mean_absolute_percentage_error, mean_absolute_error\n",
        "from xgboost import XGBRegressor\n",
        "from catboost import CatBoostRegressor"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# загрузим данные\n",
        "df = pd.read_csv('dataset.csv')"
      ],
      "metadata": {
        "id": "ukg1Ly2efMC0"
      },
      "execution_count": 159,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Описание дата-сета:**\n",
        "\n",
        "- Well_id — уникальный идентификатор скважины.\n",
        "\n",
        "- Reservoir pressure — давление жидкости в пласте, [давление].\n",
        "\n",
        "- ISIP (Instantaneous Shut-In Pressure) — давление жидкости после\n",
        "мини-ГРП5, [давление].\n",
        "\n",
        "- Closure Pressure — давление жидкости, когда трещина полностью\n",
        "     закрывается, [давление].\n",
        "\n",
        "- PAD Volume — объем чистой жидкости, закачанной в трещину, [объем].\n",
        "\n",
        "- Fluid Efficiency — эффективность жидкости гидроразрыва (вычисляется\n",
        "     как отношение объема трещины к объему закачанной жидкости), [0-1].\n",
        "\n",
        "- Transmissibility (пропускная способность) — коэффициент,\n",
        "     характеризующий сопротивление жидкости течению через пласт.\n",
        "- Total Prop Mass — общая масса закачанного проппанта во время\n",
        "     основной обработки, [масса].\n",
        "\n",
        "- Max Prop Conc — максимальная концентрация проппанта во время\n",
        "     основной обработки, [концентрация]."
      ],
      "metadata": {
        "id": "Gd3FZJkfwUjD"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "JVe2k1tjfVO3",
        "outputId": "f63bacfe-1a82-4890-f725-5785382a416f"
      },
      "execution_count": 160,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-937745a4-7121-4783-9959-d3e7e648fbc0\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Well_id</th>\n",
              "      <th>Reservoir Pressure</th>\n",
              "      <th>ISIP</th>\n",
              "      <th>Closure Pressure</th>\n",
              "      <th>PAD Volume</th>\n",
              "      <th>Fluid Efficiency</th>\n",
              "      <th>Transmissibility</th>\n",
              "      <th>Total Prop Mass</th>\n",
              "      <th>Max Prop Conc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>1250.50</td>\n",
              "      <td>2129.5</td>\n",
              "      <td>1861.0</td>\n",
              "      <td>4.892</td>\n",
              "      <td>0.149</td>\n",
              "      <td>12.685</td>\n",
              "      <td>16787.317</td>\n",
              "      <td>7.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>946.75</td>\n",
              "      <td>2850.5</td>\n",
              "      <td>2413.5</td>\n",
              "      <td>5.005</td>\n",
              "      <td>0.240</td>\n",
              "      <td>10.236</td>\n",
              "      <td>14543.000</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>1515.00</td>\n",
              "      <td>3782.5</td>\n",
              "      <td>3054.5</td>\n",
              "      <td>4.995</td>\n",
              "      <td>0.100</td>\n",
              "      <td>10.667</td>\n",
              "      <td>14366.000</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>2390.55</td>\n",
              "      <td>3449.5</td>\n",
              "      <td>3048.0</td>\n",
              "      <td>4.457</td>\n",
              "      <td>0.040</td>\n",
              "      <td>10.765</td>\n",
              "      <td>15406.000</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>8</td>\n",
              "      <td>655.95</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2290.0</td>\n",
              "      <td>3.563</td>\n",
              "      <td>0.080</td>\n",
              "      <td>12.254</td>\n",
              "      <td>15504.917</td>\n",
              "      <td>7.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>9</td>\n",
              "      <td>1837.95</td>\n",
              "      <td>4592.5</td>\n",
              "      <td>3053.0</td>\n",
              "      <td>4.988</td>\n",
              "      <td>0.360</td>\n",
              "      <td>10.404</td>\n",
              "      <td>14162.417</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>10</td>\n",
              "      <td>927.00</td>\n",
              "      <td>3503.0</td>\n",
              "      <td>2331.0</td>\n",
              "      <td>4.404</td>\n",
              "      <td>0.290</td>\n",
              "      <td>10.986</td>\n",
              "      <td>16674.917</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>11</td>\n",
              "      <td>1634.70</td>\n",
              "      <td>3758.5</td>\n",
              "      <td>2694.5</td>\n",
              "      <td>4.988</td>\n",
              "      <td>0.020</td>\n",
              "      <td>12.067</td>\n",
              "      <td>12597.417</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>12</td>\n",
              "      <td>1318.15</td>\n",
              "      <td>4253.0</td>\n",
              "      <td>3202.5</td>\n",
              "      <td>4.628</td>\n",
              "      <td>0.160</td>\n",
              "      <td>10.248</td>\n",
              "      <td>11694.583</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>13</td>\n",
              "      <td>1140.40</td>\n",
              "      <td>3062.0</td>\n",
              "      <td>2369.0</td>\n",
              "      <td>4.454</td>\n",
              "      <td>0.090</td>\n",
              "      <td>13.169</td>\n",
              "      <td>15655.500</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>15</td>\n",
              "      <td>2142.85</td>\n",
              "      <td>2850.0</td>\n",
              "      <td>2354.5</td>\n",
              "      <td>3.985</td>\n",
              "      <td>0.530</td>\n",
              "      <td>10.256</td>\n",
              "      <td>14289.500</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>16</td>\n",
              "      <td>1583.45</td>\n",
              "      <td>4348.0</td>\n",
              "      <td>3387.5</td>\n",
              "      <td>4.804</td>\n",
              "      <td>0.140</td>\n",
              "      <td>10.459</td>\n",
              "      <td>14904.083</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>17</td>\n",
              "      <td>1635.50</td>\n",
              "      <td>4494.0</td>\n",
              "      <td>3424.5</td>\n",
              "      <td>4.454</td>\n",
              "      <td>0.330</td>\n",
              "      <td>10.117</td>\n",
              "      <td>11527.317</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>18</td>\n",
              "      <td>1723.70</td>\n",
              "      <td>4947.0</td>\n",
              "      <td>3744.0</td>\n",
              "      <td>5.698</td>\n",
              "      <td>0.170</td>\n",
              "      <td>10.021</td>\n",
              "      <td>12251.100</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>19</td>\n",
              "      <td>1008.90</td>\n",
              "      <td>2668.0</td>\n",
              "      <td>2089.0</td>\n",
              "      <td>6.235</td>\n",
              "      <td>0.140</td>\n",
              "      <td>10.976</td>\n",
              "      <td>13664.583</td>\n",
              "      <td>7.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>21</td>\n",
              "      <td>1177.35</td>\n",
              "      <td>2530.5</td>\n",
              "      <td>1876.0</td>\n",
              "      <td>4.244</td>\n",
              "      <td>0.230</td>\n",
              "      <td>10.671</td>\n",
              "      <td>12120.833</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>22</td>\n",
              "      <td>1480.75</td>\n",
              "      <td>2219.5</td>\n",
              "      <td>2085.0</td>\n",
              "      <td>4.469</td>\n",
              "      <td>0.050</td>\n",
              "      <td>11.060</td>\n",
              "      <td>14359.167</td>\n",
              "      <td>6.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>23</td>\n",
              "      <td>1466.80</td>\n",
              "      <td>3988.5</td>\n",
              "      <td>2875.0</td>\n",
              "      <td>4.336</td>\n",
              "      <td>0.240</td>\n",
              "      <td>11.237</td>\n",
              "      <td>7424.833</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>24</td>\n",
              "      <td>320.50</td>\n",
              "      <td>1837.0</td>\n",
              "      <td>755.5</td>\n",
              "      <td>3.757</td>\n",
              "      <td>0.010</td>\n",
              "      <td>10.350</td>\n",
              "      <td>NaN</td>\n",
              "      <td>4.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>25</td>\n",
              "      <td>490.60</td>\n",
              "      <td>2065.5</td>\n",
              "      <td>1282.5</td>\n",
              "      <td>4.454</td>\n",
              "      <td>0.260</td>\n",
              "      <td>11.255</td>\n",
              "      <td>18889.167</td>\n",
              "      <td>7.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>26</td>\n",
              "      <td>1923.00</td>\n",
              "      <td>5345.0</td>\n",
              "      <td>4089.0</td>\n",
              "      <td>5.384</td>\n",
              "      <td>0.170</td>\n",
              "      <td>10.050</td>\n",
              "      <td>344.583</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>28</td>\n",
              "      <td>935.70</td>\n",
              "      <td>3762.5</td>\n",
              "      <td>2867.0</td>\n",
              "      <td>5.838</td>\n",
              "      <td>0.190</td>\n",
              "      <td>10.689</td>\n",
              "      <td>18122.333</td>\n",
              "      <td>6.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>30</td>\n",
              "      <td>480.85</td>\n",
              "      <td>4175.5</td>\n",
              "      <td>2183.0</td>\n",
              "      <td>4.632</td>\n",
              "      <td>0.280</td>\n",
              "      <td>10.978</td>\n",
              "      <td>15500.000</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>31</td>\n",
              "      <td>2438.40</td>\n",
              "      <td>3900.0</td>\n",
              "      <td>2779.0</td>\n",
              "      <td>3.078</td>\n",
              "      <td>0.770</td>\n",
              "      <td>10.112</td>\n",
              "      <td>14558.000</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24</th>\n",
              "      <td>32</td>\n",
              "      <td>1360.50</td>\n",
              "      <td>3707.5</td>\n",
              "      <td>2664.5</td>\n",
              "      <td>4.657</td>\n",
              "      <td>0.430</td>\n",
              "      <td>10.063</td>\n",
              "      <td>15027.517</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25</th>\n",
              "      <td>33</td>\n",
              "      <td>618.70</td>\n",
              "      <td>4475.5</td>\n",
              "      <td>3343.5</td>\n",
              "      <td>3.723</td>\n",
              "      <td>0.120</td>\n",
              "      <td>11.245</td>\n",
              "      <td>14037.000</td>\n",
              "      <td>6.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>26</th>\n",
              "      <td>34</td>\n",
              "      <td>1630.90</td>\n",
              "      <td>4766.5</td>\n",
              "      <td>3397.0</td>\n",
              "      <td>4.454</td>\n",
              "      <td>0.260</td>\n",
              "      <td>10.145</td>\n",
              "      <td>6363.083</td>\n",
              "      <td>3.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>27</th>\n",
              "      <td>35</td>\n",
              "      <td>1860.80</td>\n",
              "      <td>4322.5</td>\n",
              "      <td>2821.5</td>\n",
              "      <td>4.498</td>\n",
              "      <td>0.410</td>\n",
              "      <td>10.299</td>\n",
              "      <td>17250.000</td>\n",
              "      <td>6.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>28</th>\n",
              "      <td>36</td>\n",
              "      <td>1545.35</td>\n",
              "      <td>1948.5</td>\n",
              "      <td>1727.5</td>\n",
              "      <td>4.456</td>\n",
              "      <td>0.530</td>\n",
              "      <td>10.211</td>\n",
              "      <td>15183.333</td>\n",
              "      <td>5.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>29</th>\n",
              "      <td>38</td>\n",
              "      <td>502.00</td>\n",
              "      <td>2861.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>4.288</td>\n",
              "      <td>0.140</td>\n",
              "      <td>12.349</td>\n",
              "      <td>15476.417</td>\n",
              "      <td>9.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>30</th>\n",
              "      <td>39</td>\n",
              "      <td>2653.95</td>\n",
              "      <td>2986.5</td>\n",
              "      <td>2774.5</td>\n",
              "      <td>3.858</td>\n",
              "      <td>0.580</td>\n",
              "      <td>10.351</td>\n",
              "      <td>22828.500</td>\n",
              "      <td>8.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>31</th>\n",
              "      <td>40</td>\n",
              "      <td>257.10</td>\n",
              "      <td>2548.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.983</td>\n",
              "      <td>0.020</td>\n",
              "      <td>14.339</td>\n",
              "      <td>10247.167</td>\n",
              "      <td>8.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>32</th>\n",
              "      <td>41</td>\n",
              "      <td>106.15</td>\n",
              "      <td>2050.0</td>\n",
              "      <td>1416.5</td>\n",
              "      <td>4.298</td>\n",
              "      <td>0.100</td>\n",
              "      <td>13.340</td>\n",
              "      <td>11416.667</td>\n",
              "      <td>8.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>33</th>\n",
              "      <td>43</td>\n",
              "      <td>1487.25</td>\n",
              "      <td>3891.0</td>\n",
              "      <td>3134.0</td>\n",
              "      <td>1.598</td>\n",
              "      <td>0.100</td>\n",
              "      <td>11.330</td>\n",
              "      <td>6993.583</td>\n",
              "      <td>8.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>34</th>\n",
              "      <td>44</td>\n",
              "      <td>953.50</td>\n",
              "      <td>2651.5</td>\n",
              "      <td>2279.0</td>\n",
              "      <td>2.671</td>\n",
              "      <td>0.090</td>\n",
              "      <td>12.712</td>\n",
              "      <td>7166.667</td>\n",
              "      <td>8.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>35</th>\n",
              "      <td>45</td>\n",
              "      <td>1340.20</td>\n",
              "      <td>2400.5</td>\n",
              "      <td>1976.0</td>\n",
              "      <td>3.562</td>\n",
              "      <td>0.180</td>\n",
              "      <td>12.828</td>\n",
              "      <td>20749.583</td>\n",
              "      <td>10.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>36</th>\n",
              "      <td>46</td>\n",
              "      <td>1355.60</td>\n",
              "      <td>2514.5</td>\n",
              "      <td>2084.0</td>\n",
              "      <td>3.546</td>\n",
              "      <td>0.570</td>\n",
              "      <td>10.617</td>\n",
              "      <td>26874.000</td>\n",
              "      <td>10.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>37</th>\n",
              "      <td>47</td>\n",
              "      <td>937.80</td>\n",
              "      <td>2235.5</td>\n",
              "      <td>1789.0</td>\n",
              "      <td>3.558</td>\n",
              "      <td>0.060</td>\n",
              "      <td>14.008</td>\n",
              "      <td>NaN</td>\n",
              "      <td>10.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>38</th>\n",
              "      <td>48</td>\n",
              "      <td>1161.50</td>\n",
              "      <td>2427.0</td>\n",
              "      <td>1950.5</td>\n",
              "      <td>3.563</td>\n",
              "      <td>0.080</td>\n",
              "      <td>15.648</td>\n",
              "      <td>NaN</td>\n",
              "      <td>9.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>39</th>\n",
              "      <td>51</td>\n",
              "      <td>933.60</td>\n",
              "      <td>2706.0</td>\n",
              "      <td>2134.0</td>\n",
              "      <td>4.585</td>\n",
              "      <td>0.250</td>\n",
              "      <td>11.024</td>\n",
              "      <td>22932.500</td>\n",
              "      <td>8.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>40</th>\n",
              "      <td>52</td>\n",
              "      <td>2285.70</td>\n",
              "      <td>3196.0</td>\n",
              "      <td>2909.0</td>\n",
              "      <td>1.780</td>\n",
              "      <td>0.160</td>\n",
              "      <td>11.444</td>\n",
              "      <td>NaN</td>\n",
              "      <td>7.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>41</th>\n",
              "      <td>54</td>\n",
              "      <td>617.15</td>\n",
              "      <td>2975.5</td>\n",
              "      <td>2447.5</td>\n",
              "      <td>3.702</td>\n",
              "      <td>0.060</td>\n",
              "      <td>11.214</td>\n",
              "      <td>26720.917</td>\n",
              "      <td>10.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>42</th>\n",
              "      <td>55</td>\n",
              "      <td>1020.10</td>\n",
              "      <td>2420.5</td>\n",
              "      <td>1733.5</td>\n",
              "      <td>5.345</td>\n",
              "      <td>0.120</td>\n",
              "      <td>14.169</td>\n",
              "      <td>22975.000</td>\n",
              "      <td>9.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>43</th>\n",
              "      <td>56</td>\n",
              "      <td>1873.60</td>\n",
              "      <td>3235.5</td>\n",
              "      <td>3067.5</td>\n",
              "      <td>3.534</td>\n",
              "      <td>0.090</td>\n",
              "      <td>11.038</td>\n",
              "      <td>21381.083</td>\n",
              "      <td>8.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44</th>\n",
              "      <td>57</td>\n",
              "      <td>794.50</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2635.5</td>\n",
              "      <td>3.847</td>\n",
              "      <td>0.150</td>\n",
              "      <td>10.822</td>\n",
              "      <td>24806.833</td>\n",
              "      <td>9.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>45</th>\n",
              "      <td>58</td>\n",
              "      <td>955.25</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2499.0</td>\n",
              "      <td>3.563</td>\n",
              "      <td>0.170</td>\n",
              "      <td>11.331</td>\n",
              "      <td>21666.667</td>\n",
              "      <td>9.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>46</th>\n",
              "      <td>59</td>\n",
              "      <td>1248.85</td>\n",
              "      <td>2704.5</td>\n",
              "      <td>2321.0</td>\n",
              "      <td>5.150</td>\n",
              "      <td>0.070</td>\n",
              "      <td>13.088</td>\n",
              "      <td>NaN</td>\n",
              "      <td>8.4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>47</th>\n",
              "      <td>60</td>\n",
              "      <td>1319.40</td>\n",
              "      <td>3682.0</td>\n",
              "      <td>2989.0</td>\n",
              "      <td>3.556</td>\n",
              "      <td>0.100</td>\n",
              "      <td>11.808</td>\n",
              "      <td>26301.583</td>\n",
              "      <td>9.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-937745a4-7121-4783-9959-d3e7e648fbc0')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-937745a4-7121-4783-9959-d3e7e648fbc0 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-937745a4-7121-4783-9959-d3e7e648fbc0');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ],
            "text/plain": [
              "    Well_id  Reservoir Pressure  ...  Total Prop Mass  Max Prop Conc\n",
              "0         0             1250.50  ...        16787.317            7.0\n",
              "1         2              946.75  ...        14543.000            5.0\n",
              "2         3             1515.00  ...        14366.000            5.0\n",
              "3         4             2390.55  ...        15406.000            5.0\n",
              "4         8              655.95  ...        15504.917            7.0\n",
              "5         9             1837.95  ...        14162.417            5.0\n",
              "6        10              927.00  ...        16674.917            5.0\n",
              "7        11             1634.70  ...        12597.417            5.0\n",
              "8        12             1318.15  ...        11694.583            5.0\n",
              "9        13             1140.40  ...        15655.500            5.0\n",
              "10       15             2142.85  ...        14289.500            5.0\n",
              "11       16             1583.45  ...        14904.083            5.0\n",
              "12       17             1635.50  ...        11527.317            5.0\n",
              "13       18             1723.70  ...        12251.100            5.0\n",
              "14       19             1008.90  ...        13664.583            7.0\n",
              "15       21             1177.35  ...        12120.833            5.0\n",
              "16       22             1480.75  ...        14359.167            6.0\n",
              "17       23             1466.80  ...         7424.833            5.0\n",
              "18       24              320.50  ...              NaN            4.0\n",
              "19       25              490.60  ...        18889.167            7.0\n",
              "20       26             1923.00  ...          344.583            2.0\n",
              "21       28              935.70  ...        18122.333            6.0\n",
              "22       30              480.85  ...        15500.000            5.0\n",
              "23       31             2438.40  ...        14558.000            5.0\n",
              "24       32             1360.50  ...        15027.517            5.0\n",
              "25       33              618.70  ...        14037.000            6.0\n",
              "26       34             1630.90  ...         6363.083            3.0\n",
              "27       35             1860.80  ...        17250.000            6.0\n",
              "28       36             1545.35  ...        15183.333            5.0\n",
              "29       38              502.00  ...        15476.417            9.0\n",
              "30       39             2653.95  ...        22828.500            8.0\n",
              "31       40              257.10  ...        10247.167            8.0\n",
              "32       41              106.15  ...        11416.667            8.0\n",
              "33       43             1487.25  ...         6993.583            8.0\n",
              "34       44              953.50  ...         7166.667            8.0\n",
              "35       45             1340.20  ...        20749.583           10.0\n",
              "36       46             1355.60  ...        26874.000           10.0\n",
              "37       47              937.80  ...              NaN           10.0\n",
              "38       48             1161.50  ...              NaN            9.0\n",
              "39       51              933.60  ...        22932.500            8.0\n",
              "40       52             2285.70  ...              NaN            7.0\n",
              "41       54              617.15  ...        26720.917           10.0\n",
              "42       55             1020.10  ...        22975.000            9.0\n",
              "43       56             1873.60  ...        21381.083            8.0\n",
              "44       57              794.50  ...        24806.833            9.0\n",
              "45       58              955.25  ...        21666.667            9.0\n",
              "46       59             1248.85  ...              NaN            8.4\n",
              "47       60             1319.40  ...        26301.583            9.0\n",
              "\n",
              "[48 rows x 9 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 160
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "В целевой переменной **Total Prop Mass** есть пропуски, отложим их в тестовую выборку:"
      ],
      "metadata": {
        "id": "1B9Q61SBMlJ2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# обучающая выборка\n",
        "df_train = df[~df['Total Prop Mass'].isna()]\n",
        "\n",
        "# тестовая выборка\n",
        "df_test = df[df['Total Prop Mass'].isna()]\n",
        "\n",
        "# удалим колонку целевой переменной из тестовой выборки\n",
        "df_test.drop(['Total Prop Mass', 'Well_id'], axis=1, inplace=True)"
      ],
      "metadata": {
        "id": "7H62s8x9Kw0G",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ca26f99d-d589-484f-a63c-7f08f0afe20e"
      },
      "execution_count": 166,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/pandas/core/frame.py:4913: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame\n",
            "\n",
            "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
            "  errors=errors,\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_train.shape, df_test.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5AtF0xRzfi7S",
        "outputId": "f20ad3bb-d441-45a2-90f7-1ceb4933bf6f"
      },
      "execution_count": 167,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((43, 9), (5, 7))"
            ]
          },
          "metadata": {},
          "execution_count": 167
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_train.info()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D3-Me01_fV77",
        "outputId": "fef459b8-247a-4417-8e96-c212ba70b1e7"
      },
      "execution_count": 168,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 43 entries, 0 to 47\n",
            "Data columns (total 9 columns):\n",
            " #   Column              Non-Null Count  Dtype  \n",
            "---  ------              --------------  -----  \n",
            " 0   Well_id             43 non-null     int64  \n",
            " 1   Reservoir Pressure  43 non-null     float64\n",
            " 2   ISIP                40 non-null     float64\n",
            " 3   Closure Pressure    41 non-null     float64\n",
            " 4   PAD Volume          43 non-null     float64\n",
            " 5   Fluid Efficiency    43 non-null     float64\n",
            " 6   Transmissibility    43 non-null     float64\n",
            " 7   Total Prop Mass     43 non-null     float64\n",
            " 8   Max Prop Conc       43 non-null     float64\n",
            "dtypes: float64(8), int64(1)\n",
            "memory usage: 3.4 KB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "В датасете 48 наблюдений и 9 колонок, все в числовом формате."
      ],
      "metadata": {
        "id": "YN7E4SwmgIwp"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Задание 1: Заполни все недостающие данные."
      ],
      "metadata": {
        "id": "qve06-lTZqZu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# пропуски в датасете\n",
        "df_train.isna().mean().sort_values(ascending=False)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sOzlyhG2ff91",
        "outputId": "360dbb9b-a090-4d92-a10d-8d7384e8fd32"
      },
      "execution_count": 169,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "ISIP                  0.069767\n",
              "Closure Pressure      0.046512\n",
              "Well_id               0.000000\n",
              "Reservoir Pressure    0.000000\n",
              "PAD Volume            0.000000\n",
              "Fluid Efficiency      0.000000\n",
              "Transmissibility      0.000000\n",
              "Total Prop Mass       0.000000\n",
              "Max Prop Conc         0.000000\n",
              "dtype: float64"
            ]
          },
          "metadata": {},
          "execution_count": 169
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Видим пропуски в двух признаках:\n",
        "\n",
        "- ISIP ~ 6%;\n",
        "- Closure Pressure ~ 4%."
      ],
      "metadata": {
        "id": "KoCIkVr_j_i3"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Заполним пропуски. Для этого используем простой, но очень эффективный алгоритм: **KNN Imputer**.\n",
        "\n",
        "Для оптимизации модели, переберем в итерации несколько 'k' и на каждом шаге будем обучать XGBRegressor и прогнозировать 'Total Prop Mass'. Модель с наименьшей ошибкой и будет соответсвовать оптимальному 'k'."
      ],
      "metadata": {
        "id": "kuVGohi0kurz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def optimize_k(dataframe, target):\n",
        "  \"\"\"\n",
        "  Поиск оптимального k для KNNImputer\n",
        "\n",
        "  :param dataframe: pd.DataFrame, датафрейм, на котором будет обучаться модель\n",
        "  :param target: str, название целевой переменной\n",
        "  :return: list, список словарей с k и ошибкой модели\n",
        "  \"\"\"\n",
        "  errors = []\n",
        "  for k in range(1, 21, 2):\n",
        "    # инициализация и применение модели к данным\n",
        "    imputer = KNNImputer(n_neighbors=k)\n",
        "    imputed = imputer.fit_transform(dataframe)\n",
        "    # трансформируем данные обратно в датафрейм\n",
        "    df = pd.DataFrame(imputed, columns=dataframe.columns)\n",
        "\n",
        "    # разбиваем данные на тренировочную и валидационную выборку\n",
        "    X_train, X_val, y_train, y_val = train_test_split(df.drop(target, axis=1),\n",
        "                                                      df[target],\n",
        "                                                      random_state=23)\n",
        "    # иинциализируем xgboost и обучаем модель\n",
        "    x = XGBRegressor(random_state=23)\n",
        "    x.fit(X_train, y_train)\n",
        "    # прогноз валидационной выборки\n",
        "    pred = x.predict(X_val)\n",
        "    # ошибка модели\n",
        "    mape = mean_absolute_percentage_error(y_val, pred)\n",
        "    # добавляем в список ошибок k и соответсвующую ей ошибку\n",
        "    errors.append(({'k': k, 'MAPE': mape}))\n",
        "\n",
        "\n",
        "  return errors"
      ],
      "metadata": {
        "id": "5fPEBDCOqcDv"
      },
      "execution_count": 170,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "errors = optimize_k(df, 'Total Prop Mass')"
      ],
      "metadata": {
        "id": "aaNYG8WBs08u"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "errors"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Gr8woPftS5N2",
        "outputId": "03bc8e83-3d1c-42ed-c6c4-29a9a6aad450"
      },
      "execution_count": 172,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'MAPE': 2.1015862494749435, 'k': 1},\n",
              " {'MAPE': 1.8911976002327189, 'k': 3},\n",
              " {'MAPE': 1.9118933426625386, 'k': 5},\n",
              " {'MAPE': 1.905202177113324, 'k': 7},\n",
              " {'MAPE': 1.9678472865888514, 'k': 9},\n",
              " {'MAPE': 1.9209157603616573, 'k': 11},\n",
              " {'MAPE': 1.8680140057419512, 'k': 13},\n",
              " {'MAPE': 1.8962602771909198, 'k': 15},\n",
              " {'MAPE': 1.9113803751645448, 'k': 17},\n",
              " {'MAPE': 1.9066216787893364, 'k': 19}]"
            ]
          },
          "metadata": {},
          "execution_count": 172
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Оптимальное 'k' = 3.\n",
        "\n",
        "Заполним пропуски:"
      ],
      "metadata": {
        "id": "LyvlXaXuULXm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# иинциализация модели\n",
        "imputer = KNNImputer(n_neighbors=3)\n",
        "# применим модель\n",
        "imputed = imputer.fit_transform(df_train)\n",
        "# конвертируем в датафрейм\n",
        "df_train_imp = pd.DataFrame(imputed, columns=df_train.columns)"
      ],
      "metadata": {
        "id": "EjtsE9x1SgU_"
      },
      "execution_count": 136,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Задание 2: Проверь гипотезу о связи данных."
      ],
      "metadata": {
        "id": "KwIDJN6TZk6s"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Наш коллега, занимающийся созданием планов проведения работ по гидроразрыву пласта, однажды поделился своим эмпирическим правилом (rule of thumb): чем больше давление в резервуаре, тем больше надо закачать в трещину проппана."
      ],
      "metadata": {
        "id": "LqaUqtlUv0Oj"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "- Проверим так ли это на самом деле, применив критерий Спирмена, так как он не требует нормальности данных."
      ],
      "metadata": {
        "id": "ksRF4nBaBum-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# критерий Спирмена\n",
        "stats.spearmanr(df_train_imp['Reservoir Pressure'], df_train_imp['Total Prop Mass'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QOE_NDLe3yyk",
        "outputId": "989a2a05-b329-4666-a0aa-0edda39f5a29"
      },
      "execution_count": 137,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SpearmanrResult(correlation=-0.18317728782845064, pvalue=0.23969114977857986)"
            ]
          },
          "metadata": {},
          "execution_count": 137
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- видим, что критерий уверенно принимает нулевую гипотезу, о том, что корелляция между двумя наблюдениями незначима, следовательно взаимозависимости между давлением в резеурвуаре и количеством пропана, требуемого к закачке, **нет**."
      ],
      "metadata": {
        "id": "3KwtNmphATpe"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Посмотрим, есть ли в датасете характеристики, которые сильно взаимосвязаны:"
      ],
      "metadata": {
        "id": "Nc41hjz-CSh8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(8, 6))\n",
        "sns.heatmap(df_train_imp.drop('Well_id', axis=1).corr(method='spearman'), annot=True, linewidths=0.005, cmap='coolwarm')\n",
        "plt.title('Correlation map')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 473
        },
        "id": "GOeTkiGZV50b",
        "outputId": "a8205da0-4db7-4c6c-954d-7b4f75077113"
      },
      "execution_count": 138,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- Весьма высокая корреляция наблюдается между **ISIP**(давление жидкости после мини-ГРП) и **Closure Pressure**(давление жидкости, когда трещина полностью закрывается)\n",
        "\n",
        "Проверим ее значимость:"
      ],
      "metadata": {
        "id": "kgl7Lz3QDGgr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# критерий Спирмена\n",
        "stats.spearmanr(df_train_imp['ISIP'], df_train_imp['Closure Pressure'])"
      ],
      "metadata": {
        "id": "Ent741gCab7v",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4f9e370a-6c07-4fd7-cf8c-acb3c01793a7"
      },
      "execution_count": 139,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SpearmanrResult(correlation=0.9036544850498339, pvalue=1.0833260545810864e-16)"
            ]
          },
          "metadata": {},
          "execution_count": 139
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- критерий уверенно отвергает нулевую гипотезу о том, что корелляция между двумя наблюдениями незначима, следовательно между **ISIP** и Closure **Pressure** действительно очень высокая взаимосвязь."
      ],
      "metadata": {
        "id": "-UTQKHtQOyz9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Задание 3: Натренируй модель машинного обучения."
      ],
      "metadata": {
        "id": "Qlx24E5oQjGn"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Опираясь на имеющиеся данные, построй модель для предсказания массы проппанта **Total\n",
        "Prop Mass** и оцени ее точность.\n"
      ],
      "metadata": {
        "id": "hYaAgck_QmXM"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Обучим XGBRegressor и СatBoostRegressor и усредним их прогноз."
      ],
      "metadata": {
        "id": "4N86ueKgc9YK"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Проверим, нет ли в данных дупликатов:"
      ],
      "metadata": {
        "id": "33d0NmbdRN42"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "duplicates = df_train_imp.duplicated().sum()\n",
        "\n",
        "print(f'В датасете {duplicates} дупликатов.')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-NOP2FwNOjbE",
        "outputId": "302aa4a9-2a14-425a-e26e-08876b910b60"
      },
      "execution_count": 140,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "В датасете 0 дупликатов.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Посмотрим на распределение целевой переменной:"
      ],
      "metadata": {
        "id": "NWwr0u7ZXdKO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "sns.histplot(df_train_imp['Total Prop Mass'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "IwtqxVO_SnZU",
        "outputId": "301fd1de-2a97-4577-d6ae-d850e93ef37d"
      },
      "execution_count": 141,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- распределение целевой переменной не похоже на нормальное, проверим это, применив критерии Шапиро-Уилка и Харке-Бера:"
      ],
      "metadata": {
        "id": "Yvz_nAoKXhg3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# критерий Шапиро-Уилка\n",
        "stat, p_value = stats.shapiro(df_train_imp['Total Prop Mass'])\n",
        "\n",
        "print(f'Статистика: {stat}, p_value: {p_value}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ha0hYTzCXsEG",
        "outputId": "f6c51b95-d3e2-4096-edbf-09b5fef23998"
      },
      "execution_count": 142,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Статистика: 0.96145099401474, p_value: 0.15664900839328766\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# критерий Харке-Бера\n",
        "stat, p_value = stats.jarque_bera(df_train_imp['Total Prop Mass'])\n",
        "\n",
        "print(f'Статистика: {stat}, p_value: {p_value}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vW21KsphYu_W",
        "outputId": "b9ff1cb1-9d79-4c79-d07a-906e5b121948"
      },
      "execution_count": 143,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Статистика: 0.05937375660952339, p_value: 0.9707494486778364\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- Оба теста уверенно приняли нулевую гипотезу, что распределение целевой переменной - нормальное, поэтому мы не будем ее логарифмировать для обучения модели."
      ],
      "metadata": {
        "id": "S6qNu205ZDtW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# разобьем датасет на обучающую и валидационную выборки\n",
        "train = df_train_imp.drop(['Well_id', 'Total Prop Mass'], axis=1)\n",
        "y = df_train_imp['Total Prop Mass']\n",
        "\n",
        "X_train, X_val, y_train, y_val = train_test_split(train, y, train_size=0.8, random_state=123)"
      ],
      "metadata": {
        "id": "3dhnxiK9TARt"
      },
      "execution_count": 144,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train.shape, X_val.shape, y_train.shape, y_val.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "S2ZxYSQzVKHA",
        "outputId": "fb16b305-ec70-4fb3-8da2-9694abf18a36"
      },
      "execution_count": 145,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((34, 7), (9, 7), (34,), (9,))"
            ]
          },
          "metadata": {},
          "execution_count": 145
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### XGBRegressor:"
      ],
      "metadata": {
        "id": "wh5j3Vaybn15"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# инициализация\n",
        "x = XGBRegressor(n_estimators=300, max_depth=8, random_state=23)\n",
        "\n",
        "# обучение\n",
        "x.fit(X_train, y_train)\n",
        "\n",
        "# прогноз валидационной выборки\n",
        "xgb_predict_val = x.predict(X_val)\n",
        "\n",
        "# оценка модели\n",
        "xgb_mape = mean_absolute_percentage_error(y_val, xgb_predict_val)\n",
        "xgb_mae = mean_absolute_error(y_val, xgb_predict_val)\n",
        "\n",
        "print(f'\\nMAPE xgboost на валидационной выборке: {xgb_mape}')\n",
        "print(f'MAE xgboost на валидационной выборке: {xgb_mae}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8-gDmSaGVNPI",
        "outputId": "4c0e61d4-cf09-4ec6-86f4-896c33372ffe"
      },
      "execution_count": 146,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[10:06:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n",
            "\n",
            "MAPE xgboost на валидационной выборке: 0.1187532120473023\n",
            "MAE xgboost на валидационной выборке: 1810.1937925347224\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Видим, что XGBRegressor в среднем ошибается на 11.87%, что неплохой результат."
      ],
      "metadata": {
        "id": "Gg0rciVgghIQ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### CatBoostRegressor:"
      ],
      "metadata": {
        "id": "kEGKHoSBeAVF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# инициализация\n",
        "cb = CatBoostRegressor(iterations=170, max_depth=8,  random_state=23, verbose=False)\n",
        "\n",
        "# обучение\n",
        "cb.fit(X_train, y_train)\n",
        "\n",
        "# прогноз валидационной выборки\n",
        "cb_predict_val = cb.predict(X_val)\n",
        "\n",
        "# оценка модели\n",
        "cb_mape = mean_absolute_percentage_error(y_val, cb_predict_val)\n",
        "cb_mae = mean_absolute_error(y_val, cb_predict_val)\n",
        "\n",
        "print(f'MAPE catboost на валидационной выборке: {cb_mape}')\n",
        "print(f'MAE catboost на валидационной выборке: {cb_mae}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EWaDy-M6eJ5u",
        "outputId": "36ace891-5e8d-45c3-b0d2-2b8e14aba051"
      },
      "execution_count": 147,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MAPE catboost на валидационной выборке: 0.09908682744287606\n",
            "MAE catboost на валидационной выборке: 1451.989876578883\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "CatBoostRegressor в среднем ошибается на 9.90%, что почти на 2% лучше чем XGBRegressor."
      ],
      "metadata": {
        "id": "4PXMOFIHhE1M"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Объединим два алгоритма:"
      ],
      "metadata": {
        "id": "HPxvcng6hVYU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# объединение двух прогнозов\n",
        "total_predict_val = (xgb_predict_val + cb_predict_val) / 2\n",
        "\n",
        "# MAPE после объединения\n",
        "total_mape = mean_absolute_percentage_error(y_val, total_predict_val)\n",
        "\n",
        "print(f'MAPE после объединения двух алгоритмов: {total_mape}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YoNwwgbdeN2n",
        "outputId": "d81f0847-45d8-47c0-d7e0-3863d98a70c2"
      },
      "execution_count": 148,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MAPE после объединения двух алгоритмов: 0.0913265959434874\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Объединение двух алгоритмов позволило **улучшить** метрику на валидационной выборке еще на 0.77%."
      ],
      "metadata": {
        "id": "ERmhFOlZiyu4"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Прогноз тестовой выборки:"
      ],
      "metadata": {
        "id": "8by3tJfojteV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# xgboostregressor\n",
        "xgb_predict_test = x.predict(df_test)\n",
        "\n",
        "# catboostregressor\n",
        "cb_predict_test = cb.predict(df_test)\n",
        "\n",
        "# объединенный прогноз\n",
        "test_prediction = (xgb_predict_test + cb_predict_test) / 2\n",
        "\n",
        "print(test_prediction)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UrBUuRRbXfFT",
        "outputId": "99d20c8b-17ff-42e3-fd7d-90b03f143cd8"
      },
      "execution_count": 149,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[15134.70724829 20686.03059998 21160.60976884 11831.81973288\n",
            " 18100.03571718]\n"
          ]
        }
      ]
    }
  ]
}