Taxi_trip_duration_Anh_NGUYEN.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Taxi-trip-duration.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "metadata": {
        "id": "Xwv0TSTwJu9k",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Import module and installation package"
      ]
    },
    {
      "metadata": {
        "id": "3niTJoyfS6gF",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import pandas as pd"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "0df99kXWWtt-",
        "colab_type": "code",
        "outputId": "4ba2e93e-3609-477c-f172-3847eeab5088",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "from googleapiclient.discovery import build\n",
        "import io, os\n",
        "from googleapiclient.http import MediaIoBaseDownload\n",
        "from google.colab import auth\n",
        "\n",
        "auth.authenticate_user()\n",
        "\n",
        "drive_service = build('drive', 'v3')\n",
        "# be sure your API key file json copied to your drive google\n",
        "results = drive_service.files().list(q=\"name = 'kaggle.json'\", fields=\"files(id)\").execute()\n",
        "kaggle_api_key = results.get('files', [])\n",
        "\n",
        "filename = \"/root/.kaggle/kaggle.json\"\n",
        "os.makedirs(os.path.dirname(filename), exist_ok=True)\n",
        "\n",
        "request = drive_service.files().get_media(fileId=kaggle_api_key[0]['id'])\n",
        "fh = io.FileIO(filename, 'wb')\n",
        "downloader = MediaIoBaseDownload(fh, request)\n",
        "done = False\n",
        "while done is False:\n",
        "    status, done = downloader.next_chunk()\n",
        "    print(\"Download %d%%.\" %int(status.progress()*100))\n",
        "\n",
        "os.chmod(filename, 600)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Download 100%.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "oQcc2vKuVm3x",
        "colab_type": "code",
        "outputId": "faedaff0-0212-43da-eee4-c725e4c2f893",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 361
        }
      },
      "cell_type": "code",
      "source": [
        "!pip install kaggle;\n",
        "!pip install lightgbm;\n",
        "!pip install xgboost;"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: kaggle in /usr/local/lib/python3.6/dist-packages (1.5.3)\n",
            "Requirement already satisfied: urllib3<1.25,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from kaggle) (1.22)\n",
            "Requirement already satisfied: six>=1.10 in /usr/local/lib/python3.6/dist-packages (from kaggle) (1.11.0)\n",
            "Requirement already satisfied: certifi in /usr/local/lib/python3.6/dist-packages (from kaggle) (2018.11.29)\n",
            "Requirement already satisfied: python-dateutil in /usr/local/lib/python3.6/dist-packages (from kaggle) (2.5.3)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from kaggle) (2.18.4)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.6/dist-packages (from kaggle) (4.28.1)\n",
            "Requirement already satisfied: python-slugify in /usr/local/lib/python3.6/dist-packages (from kaggle) (2.0.1)\n",
            "Requirement already satisfied: chardet<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->kaggle) (3.0.4)\n",
            "Requirement already satisfied: idna<2.7,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->kaggle) (2.6)\n",
            "Requirement already satisfied: Unidecode>=0.04.16 in /usr/local/lib/python3.6/dist-packages (from python-slugify->kaggle) (1.0.23)\n",
            "Requirement already satisfied: lightgbm in /usr/local/lib/python3.6/dist-packages (2.2.3)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from lightgbm) (1.14.6)\n",
            "Requirement already satisfied: scikit-learn in /usr/local/lib/python3.6/dist-packages (from lightgbm) (0.20.2)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.6/dist-packages (from lightgbm) (1.1.0)\n",
            "Requirement already satisfied: xgboost in /usr/local/lib/python3.6/dist-packages (0.7.post4)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from xgboost) (1.14.6)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.6/dist-packages (from xgboost) (1.1.0)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "pIZKxSzZZ0Lr",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from scipy import stats\n",
        "from scipy.stats import norm, skew\n",
        "from sklearn.preprocessing import power_transform"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "aXanLoDpav1L",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "sns.set_style(\"darkgrid\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "3bNchcb8LrSA",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "import lightgbm as lgb\n",
        "import xgboost as xgb"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "lk9MI_W2dFRE",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from sklearn.tree import DecisionTreeRegressor, ExtraTreeRegressor\n",
        "from sklearn.linear_model import ElasticNet, Lasso, Ridge, SGDRegressor\n",
        "from sklearn.ensemble import AdaBoostRegressor, BaggingRegressor, ExtraTreesRegressor, GradientBoostingRegressor, RandomForestRegressor\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "P4NtBrumGbFI",
        "colab_type": "code",
        "outputId": "33aed5cb-204e-4ba4-ca38-b1c61809f40a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        }
      },
      "cell_type": "code",
      "source": [
        "# download output file from kernel\n",
        "!kaggle kernels output nyc-taxi-duration-eda-by-nguyen-khac-bao-anh"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Output file downloaded to /content/testing_data.csv\n",
            "Output file downloaded to /content/training_data.csv\n",
            "Kernel log downloaded to /content/nyc-taxi-duration-eda-by-nguyen-khac-bao-anh.log \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "O7ejHsa4J6hT",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Data reading"
      ]
    },
    {
      "metadata": {
        "id": "u8G35xggI4UT",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "## read data\n",
        "train = pd.read_csv('training_data.csv')\n",
        "test = pd.read_csv('testing_data.csv')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "L-3qFhbGlikw",
        "colab_type": "code",
        "outputId": "382faa26-afde-4c51-9413-b747875147ba",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "col_have_nan = list(set(train.columns[train.isna().sum() != 0]).union(set(test.columns[test.isna().sum() != 0])))\n",
        "col_have_nan"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['ave_speed_x', 'Temp.', 'ave_speed_y']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 104
        }
      ]
    },
    {
      "metadata": {
        "id": "Js9BJnTDmuRi",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "for c in col_have_nan:\n",
        "    train[c].fillna(train[c].mean(), inplace=True)\n",
        "    test[c].fillna(train[c].mean(), inplace=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "wA6OvEXNYalT",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "xtrain = train.drop(['id', 'pickup_datetime', 'dropoff_datetime', 'trip_duration', 'log_trip_duration'], axis=1)\n",
        "xtest = test.drop(['id', 'pickup_datetime'], axis=1)\n",
        "y = train.log_trip_duration.values\n",
        "id_test = test.id.values"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "eGee5n2PZoa7",
        "colab_type": "code",
        "outputId": "fa8b17fb-d03e-4d6e-8291-21b4a544b793",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "xtrain.shape, xtest.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((1458644, 21), (625134, 21))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 111
        }
      ]
    },
    {
      "metadata": {
        "id": "wFj_0sNlZE9u",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Setup for all data like cox-box transform"
      ]
    },
    {
      "metadata": {
        "id": "aYzvKvZGZWwy",
        "colab_type": "code",
        "outputId": "f3558f4c-d237-491c-b9c6-a635b7df8e66",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "#ntrain = train.shape[0]\n",
        "#ntest = test.shape[0]\n",
        "#id_test = test.id.values\n",
        "#y = train.log_trip_duration.values\n",
        "#all_data = pd.concat([train, test])\n",
        "#all_data.drop(['id', 'pickup_datetime', 'dropoff_datetime', 'trip_duration', 'log_trip_duration'], axis=1, inplace=True)\n",
        "#del(train, test)\n",
        "#print(\"all_data size is : {}\".format(all_data.shape))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "all_data size is : (2083778, 21)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "lPO6zKXPbAnQ",
        "colab_type": "code",
        "outputId": "014fb91c-b7f8-453f-e95b-051e65425c9d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 253
        }
      },
      "cell_type": "code",
      "source": [
        "#all_data.head()"
      ],
      "execution_count": 0,
      "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>0_x</th>\n",
              "      <th>0_y</th>\n",
              "      <th>Temp.</th>\n",
              "      <th>ave_speed_x</th>\n",
              "      <th>ave_speed_y</th>\n",
              "      <th>day</th>\n",
              "      <th>dayofweek</th>\n",
              "      <th>dist</th>\n",
              "      <th>dist_lat</th>\n",
              "      <th>dist_long</th>\n",
              "      <th>...</th>\n",
              "      <th>dropoff_longitude</th>\n",
              "      <th>hour</th>\n",
              "      <th>minute</th>\n",
              "      <th>month</th>\n",
              "      <th>passenger_count</th>\n",
              "      <th>pickup_latitude</th>\n",
              "      <th>pickup_longitude</th>\n",
              "      <th>store_and_fwd_flag</th>\n",
              "      <th>vendor_id</th>\n",
              "      <th>weekday</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>68221</td>\n",
              "      <td>72764</td>\n",
              "      <td>4.4</td>\n",
              "      <td>3.610834</td>\n",
              "      <td>3.898475</td>\n",
              "      <td>14</td>\n",
              "      <td>0</td>\n",
              "      <td>0.017680</td>\n",
              "      <td>0.002335</td>\n",
              "      <td>-0.017525</td>\n",
              "      <td>...</td>\n",
              "      <td>-73.964630</td>\n",
              "      <td>17</td>\n",
              "      <td>24</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>40.767937</td>\n",
              "      <td>-73.982155</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>59478</td>\n",
              "      <td>51890</td>\n",
              "      <td>28.9</td>\n",
              "      <td>3.855183</td>\n",
              "      <td>3.548049</td>\n",
              "      <td>12</td>\n",
              "      <td>6</td>\n",
              "      <td>0.020456</td>\n",
              "      <td>0.007412</td>\n",
              "      <td>0.019066</td>\n",
              "      <td>...</td>\n",
              "      <td>-73.999481</td>\n",
              "      <td>0</td>\n",
              "      <td>43</td>\n",
              "      <td>6</td>\n",
              "      <td>1</td>\n",
              "      <td>40.738564</td>\n",
              "      <td>-73.980415</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>6</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>111200</td>\n",
              "      <td>38071</td>\n",
              "      <td>-6.7</td>\n",
              "      <td>3.474121</td>\n",
              "      <td>4.320430</td>\n",
              "      <td>19</td>\n",
              "      <td>1</td>\n",
              "      <td>0.059934</td>\n",
              "      <td>0.053852</td>\n",
              "      <td>0.026306</td>\n",
              "      <td>...</td>\n",
              "      <td>-74.005333</td>\n",
              "      <td>11</td>\n",
              "      <td>35</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>40.763939</td>\n",
              "      <td>-73.979027</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>29991</td>\n",
              "      <td>38071</td>\n",
              "      <td>7.2</td>\n",
              "      <td>4.085823</td>\n",
              "      <td>4.320430</td>\n",
              "      <td>6</td>\n",
              "      <td>2</td>\n",
              "      <td>0.013438</td>\n",
              "      <td>0.013252</td>\n",
              "      <td>0.002228</td>\n",
              "      <td>...</td>\n",
              "      <td>-74.012268</td>\n",
              "      <td>19</td>\n",
              "      <td>32</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>40.719971</td>\n",
              "      <td>-74.010040</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>36415</td>\n",
              "      <td>11401</td>\n",
              "      <td>9.4</td>\n",
              "      <td>4.407767</td>\n",
              "      <td>4.054702</td>\n",
              "      <td>26</td>\n",
              "      <td>5</td>\n",
              "      <td>0.010690</td>\n",
              "      <td>0.010689</td>\n",
              "      <td>-0.000130</td>\n",
              "      <td>...</td>\n",
              "      <td>-73.972923</td>\n",
              "      <td>13</td>\n",
              "      <td>30</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>40.793209</td>\n",
              "      <td>-73.973053</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>5</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 21 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "      0_x    0_y  Temp.  ave_speed_x  ave_speed_y  day  dayofweek      dist  \\\n",
              "0   68221  72764    4.4     3.610834     3.898475   14          0  0.017680   \n",
              "1   59478  51890   28.9     3.855183     3.548049   12          6  0.020456   \n",
              "2  111200  38071   -6.7     3.474121     4.320430   19          1  0.059934   \n",
              "3   29991  38071    7.2     4.085823     4.320430    6          2  0.013438   \n",
              "4   36415  11401    9.4     4.407767     4.054702   26          5  0.010690   \n",
              "\n",
              "   dist_lat  dist_long   ...     dropoff_longitude  hour  minute  month  \\\n",
              "0  0.002335  -0.017525   ...            -73.964630    17      24      3   \n",
              "1  0.007412   0.019066   ...            -73.999481     0      43      6   \n",
              "2  0.053852   0.026306   ...            -74.005333    11      35      1   \n",
              "3  0.013252   0.002228   ...            -74.012268    19      32      4   \n",
              "4  0.010689  -0.000130   ...            -73.972923    13      30      3   \n",
              "\n",
              "   passenger_count  pickup_latitude  pickup_longitude  store_and_fwd_flag  \\\n",
              "0                1        40.767937        -73.982155                   0   \n",
              "1                1        40.738564        -73.980415                   0   \n",
              "2                1        40.763939        -73.979027                   0   \n",
              "3                1        40.719971        -74.010040                   0   \n",
              "4                1        40.793209        -73.973053                   0   \n",
              "\n",
              "   vendor_id  weekday  \n",
              "0          2        0  \n",
              "1          1        6  \n",
              "2          2        1  \n",
              "3          2        2  \n",
              "4          2        5  \n",
              "\n",
              "[5 rows x 21 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 60
        }
      ]
    },
    {
      "metadata": {
        "id": "s-i5RDWbR9e_",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Analysis target: trip_duration"
      ]
    },
    {
      "metadata": {
        "id": "jEl1-Q6IT7LL",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Trip duration"
      ]
    },
    {
      "metadata": {
        "id": "xm6S3NdOdJfG",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "Y_transform = train.trip_duration.values"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "3ZQb2KuNSIYx",
        "colab_type": "code",
        "outputId": "d57fdd6a-2538-4ccd-cb54-8973cf9ed5a1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 829
        }
      },
      "cell_type": "code",
      "source": [
        "sns.distplot(Y_transform , fit=norm);\n",
        "\n",
        "# Get the fitted parameters used by the function\n",
        "(mu, sigma) = norm.fit(Y_transform)\n",
        "print( '\\n mu = {:.2f} and sigma = {:.2f}\\n'.format(mu, sigma))\n",
        "\n",
        "#Now plot the distribution\n",
        "plt.legend(['Normal dist. ($\\mu=$ {:.2f} and $\\sigma=$ {:.2f} )'.format(mu, sigma)],\n",
        "            loc='best')\n",
        "plt.ylabel('Frequency')\n",
        "plt.title('Trip duration distribution')\n",
        "\n",
        "#Get also the QQ-plot\n",
        "fig = plt.figure()\n",
        "res = stats.probplot(Y_transform, plot=plt)\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/matplotlib/axes/_axes.py:6521: MatplotlibDeprecationWarning: \n",
            "The 'normed' kwarg was deprecated in Matplotlib 2.1 and will be removed in 3.1. Use 'density' instead.\n",
            "  alternative=\"'density'\", removal=\"3.1\")\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "\n",
            " mu = 959.49 and sigma = 5237.43\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgYAAAFZCAYAAADjD+cfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgkAAAFnCAYAAAAsdNpPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCK0AvZq9VwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "2zbMX4v3SIiX",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Trip duration in log transform"
      ]
    },
    {
      "metadata": {
        "id": "5B8AzkfCUJkw",
        "colab_type": "code",
        "outputId": "8457f4a5-d008-464e-f5a4-1add3e0250cb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 829
        }
      },
      "cell_type": "code",
      "source": [
        "Y_transform = train['log_trip_duration'].values\n",
        "sns.distplot(Y_transform , fit=norm);\n",
        "\n",
        "# Get the fitted parameters used by the function\n",
        "(mu, sigma) = norm.fit(Y_transform)\n",
        "print( '\\n mu = {:.2f} and sigma = {:.2f}\\n'.format(mu, sigma))\n",
        "\n",
        "#Now plot the distribution\n",
        "plt.legend(['Normal dist. ($\\mu=$ {:.2f} and $\\sigma=$ {:.2f} )'.format(mu, sigma)],\n",
        "            loc='best')\n",
        "plt.ylabel('Frequency')\n",
        "plt.title('Trip duration distribution')\n",
        "\n",
        "#Get also the QQ-plot\n",
        "fig = plt.figure()\n",
        "res = stats.probplot(Y_transform, plot=plt)\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/matplotlib/axes/_axes.py:6521: MatplotlibDeprecationWarning: \n",
            "The 'normed' kwarg was deprecated in Matplotlib 2.1 and will be removed in 3.1. Use 'density' instead.\n",
            "  alternative=\"'density'\", removal=\"3.1\")\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "\n",
            " mu = 6.46 and sigma = 0.80\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFZCAYAAACBlraEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nfkIUdUEQBEHwE6KoC4IgCIKf+H+GLUKaQPE16gAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe0AAAFnCAYAAACLnxFFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "gLxmoX9BdTdX",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Test skew of features"
      ]
    },
    {
      "metadata": {
        "id": "p8ms-q_CVfPS",
        "colab_type": "code",
        "outputId": "ef824b38-ec16-4c3e-99fb-e659696b7df8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 754
        }
      },
      "cell_type": "code",
      "source": [
        "numeric_feats = all_data.dtypes[all_data.dtypes != \"object\"].index\n",
        "\n",
        "# Check the skew of all numerical features\n",
        "skewed_feats = all_data[numeric_feats].apply(lambda x: skew(x.dropna())).sort_values(ascending=False)\n",
        "print(\"\\nSkew in numerical features: \\n\")\n",
        "skewness = pd.DataFrame({'Skew' :skewed_feats})\n",
        "skewness"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Skew in numerical features: \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "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>Skew</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>ave_speed_x</th>\n",
              "      <td>580.420795</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ave_speed_y</th>\n",
              "      <td>469.961762</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dist</th>\n",
              "      <td>26.288577</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>store_and_fwd_flag</th>\n",
              "      <td>13.364071</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dist_lat</th>\n",
              "      <td>8.370125</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dist_long</th>\n",
              "      <td>5.257774</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>pickup_latitude</th>\n",
              "      <td>2.131394</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>passenger_count</th>\n",
              "      <td>2.130139</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0_y</th>\n",
              "      <td>0.320090</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0_x</th>\n",
              "      <td>0.244715</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>day</th>\n",
              "      <td>0.040479</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>minute</th>\n",
              "      <td>-0.008029</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Temp.</th>\n",
              "      <td>-0.013163</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>month</th>\n",
              "      <td>-0.015893</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>weekday</th>\n",
              "      <td>-0.042597</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dayofweek</th>\n",
              "      <td>-0.042597</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>vendor_id</th>\n",
              "      <td>-0.140063</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hour</th>\n",
              "      <td>-0.444902</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dropoff_latitude</th>\n",
              "      <td>-10.436058</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>pickup_longitude</th>\n",
              "      <td>-427.465040</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dropoff_longitude</th>\n",
              "      <td>-436.891048</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                          Skew\n",
              "ave_speed_x         580.420795\n",
              "ave_speed_y         469.961762\n",
              "dist                 26.288577\n",
              "store_and_fwd_flag   13.364071\n",
              "dist_lat              8.370125\n",
              "dist_long             5.257774\n",
              "pickup_latitude       2.131394\n",
              "passenger_count       2.130139\n",
              "0_y                   0.320090\n",
              "0_x                   0.244715\n",
              "day                   0.040479\n",
              "minute               -0.008029\n",
              "Temp.                -0.013163\n",
              "month                -0.015893\n",
              "weekday              -0.042597\n",
              "dayofweek            -0.042597\n",
              "vendor_id            -0.140063\n",
              "hour                 -0.444902\n",
              "dropoff_latitude    -10.436058\n",
              "pickup_longitude   -427.465040\n",
              "dropoff_longitude  -436.891048"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        }
      ]
    },
    {
      "metadata": {
        "id": "DM3FOa8iW3Q-",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "#### Box Cox Transformation of (highly) skewed features"
      ]
    },
    {
      "metadata": {
        "id": "juC-DZHAWzci",
        "colab_type": "code",
        "outputId": "ecd0a2a4-cff1-4390-c48d-c38b5190a24c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "skewness = skewness[abs(skewness) > 0.75]\n",
        "print(\"There are {} skewed numerical features to Box Cox transform\".format(skewness.shape[0]))\n",
        "\n",
        "from scipy.special import boxcox1p\n",
        "skewed_features = skewness.index\n",
        "lam = 0.15\n",
        "for feat in skewed_features:\n",
        "    all_data[feat] += 1\n",
        "    all_data[feat] = boxcox1p(all_data[feat], lam)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "There are 21 skewed numerical features to Box Cox transform\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "cfH8JNY7dppg",
        "colab_type": "code",
        "outputId": "8f30622a-c99c-4f74-e254-81fb2335d62b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 754
        }
      },
      "cell_type": "code",
      "source": [
        "numeric_feats = all_data.dtypes[all_data.dtypes != \"object\"].index\n",
        "\n",
        "# Check the skew of all numerical features\n",
        "skewed_feats = all_data[numeric_feats].apply(lambda x: skew(x.dropna())).sort_values(ascending=False)\n",
        "print(\"\\nSkew in numerical features: \\n\")\n",
        "skewness = pd.DataFrame({'Skew' :skewed_feats})\n",
        "skewness"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Skew in numerical features: \n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "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>Skew</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>store_and_fwd_flag</th>\n",
              "      <td>13.364071</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dist</th>\n",
              "      <td>5.805183</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ave_speed_x</th>\n",
              "      <td>3.396248</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ave_speed_y</th>\n",
              "      <td>2.593857</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>passenger_count</th>\n",
              "      <td>1.832126</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dist_long</th>\n",
              "      <td>0.770187</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>vendor_id</th>\n",
              "      <td>-0.140063</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>month</th>\n",
              "      <td>-0.339211</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dayofweek</th>\n",
              "      <td>-0.499752</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>weekday</th>\n",
              "      <td>-0.499752</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>day</th>\n",
              "      <td>-0.674244</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>minute</th>\n",
              "      <td>-0.942597</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Temp.</th>\n",
              "      <td>-1.079405</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dist_lat</th>\n",
              "      <td>-1.244161</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0_y</th>\n",
              "      <td>-1.327020</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hour</th>\n",
              "      <td>-1.428462</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0_x</th>\n",
              "      <td>-1.552371</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>pickup_latitude</th>\n",
              "      <td>-6.666834</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dropoff_latitude</th>\n",
              "      <td>-15.382984</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dropoff_longitude</th>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>pickup_longitude</th>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                         Skew\n",
              "store_and_fwd_flag  13.364071\n",
              "dist                 5.805183\n",
              "ave_speed_x          3.396248\n",
              "ave_speed_y          2.593857\n",
              "passenger_count      1.832126\n",
              "dist_long            0.770187\n",
              "vendor_id           -0.140063\n",
              "month               -0.339211\n",
              "dayofweek           -0.499752\n",
              "weekday             -0.499752\n",
              "day                 -0.674244\n",
              "minute              -0.942597\n",
              "Temp.               -1.079405\n",
              "dist_lat            -1.244161\n",
              "0_y                 -1.327020\n",
              "hour                -1.428462\n",
              "0_x                 -1.552371\n",
              "pickup_latitude     -6.666834\n",
              "dropoff_latitude   -15.382984\n",
              "dropoff_longitude         NaN\n",
              "pickup_longitude          NaN"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 51
        }
      ]
    },
    {
      "metadata": {
        "id": "WbZu7wYId3r9",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "xtrain = all_data[:ntrain]\n",
        "xtest = all_data[ntrain:]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "otpS1DclarnF",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Split data"
      ]
    },
    {
      "metadata": {
        "id": "Z8tejOl6gV0P",
        "colab_type": "code",
        "outputId": "85bb6cf7-08b8-496b-8c1a-42c904461111",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_valid, y_train, y_valid = train_test_split(xtrain,y, test_size=0.5, random_state=42)\n",
        "X_train.shape, X_valid.shape, y_train.shape, y_valid.shape"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((729322, 21), (729322, 21), (729322,), (729322,))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 112
        }
      ]
    },
    {
      "metadata": {
        "id": "XaHOUpOpcuT4",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Using `Tree`\n"
      ]
    },
    {
      "metadata": {
        "id": "wONbBHaxdA1S",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "dtr = DecisionTreeRegressor(criterion=\"mse\",max_depth=14, max_features=0.5, min_samples_split=0.5, random_state=42)\n",
        "etr = ExtraTreeRegressor(criterion=\"mse\",max_depth=14, max_features=0.5, min_samples_split=0.5, random_state=42)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "zbxo6bhDgxU2",
        "colab_type": "code",
        "outputId": "36d0c8a5-1cfe-4b5f-a9a9-14674b087be6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 107
        }
      },
      "cell_type": "code",
      "source": [
        "dtr.fit(X_train,y_train)\n",
        "etr.fit(X_train,y_train)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "ExtraTreeRegressor(criterion='mse', max_depth=14, max_features=0.5,\n",
              "          max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
              "          min_impurity_split=None, min_samples_leaf=1,\n",
              "          min_samples_split=0.5, min_weight_fraction_leaf=0.0,\n",
              "          random_state=42, splitter='random')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 114
        }
      ]
    },
    {
      "metadata": {
        "id": "Is9Ag42DpO6d",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import mean_squared_log_error\n",
        "def rmsle(y, pred):\n",
        "    return np.sqrt(mean_squared_log_error(y,pred))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "MmBiTO7dpiUD",
        "colab_type": "code",
        "outputId": "3526f9a8-2a8f-48e7-f63b-2d36acc9cda7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "cell_type": "code",
      "source": [
        "y_pred_dtr = dtr.predict(X_valid)\n",
        "y_pred_etr = etr.predict(X_valid)\n",
        "print(\"rmsle for DTR: {} \\n rmsle for ETR: {}\".format(rmsle(y_valid, y_pred_dtr),rmsle(y_valid, y_pred_etr)))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "rmsle for DTR: 0.08787277716383204 \n",
            " rmsle for ETR: 0.11349226885659831\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "jEJ4AgmbtDUg",
        "colab_type": "code",
        "outputId": "62dcbbc8-b034-470b-d59d-4254f20c6872",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 107
        }
      },
      "cell_type": "code",
      "source": [
        "dtr.feature_importances_"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.        , 0.        , 0.        , 0.        , 0.        ,\n",
              "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
              "       0.        , 0.        , 0.        , 0.61900474, 0.        ,\n",
              "       0.38099526, 0.        , 0.        , 0.        , 0.        ,\n",
              "       0.        ])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 117
        }
      ]
    },
    {
      "metadata": {
        "id": "J1vnEQgAtP8b",
        "colab_type": "code",
        "outputId": "539eadfd-b7d6-4392-af6e-506a2c0cbb4c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 107
        }
      },
      "cell_type": "code",
      "source": [
        "etr.feature_importances_"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.        , 0.        , 0.        , 0.        , 0.        ,\n",
              "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
              "       0.        , 0.        , 0.        , 0.        , 0.        ,\n",
              "       0.        , 0.17384434, 0.82615566, 0.        , 0.        ,\n",
              "       0.        ])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 118
        }
      ]
    },
    {
      "metadata": {
        "id": "3UemQacqcWfz",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "y_test_dtr = dtr.predict(xtest)\n",
        "y_test_etr = etr.predict(xtest)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "6MWetum-e_OV",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "df_train = pd.DataFrame({'dtr':y_pred_dtr, 'etr':y_pred_etr})\n",
        "df_test = pd.DataFrame({'dtr':y_test_dtr, 'etr':y_test_etr})"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "x8Ob9xvPscVX",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Using ensemble"
      ]
    },
    {
      "metadata": {
        "id": "GDXDpT5sdBZ-",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "gbr = GradientBoostingRegressor(learning_rate=0.5, n_estimators=20,subsample=0.7, max_features=0.9)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "kJVou2mftnEx",
        "colab_type": "code",
        "outputId": "23948893-9946-4122-a47c-9e22c64444b8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 161
        }
      },
      "cell_type": "code",
      "source": [
        "gbr.fit(X_train, y_train)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GradientBoostingRegressor(alpha=0.9, criterion='friedman_mse', init=None,\n",
              "             learning_rate=0.5, loss='ls', max_depth=3, max_features=0.9,\n",
              "             max_leaf_nodes=None, min_impurity_decrease=0.0,\n",
              "             min_impurity_split=None, min_samples_leaf=1,\n",
              "             min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
              "             n_estimators=20, n_iter_no_change=None, presort='auto',\n",
              "             random_state=None, subsample=0.7, tol=0.0001,\n",
              "             validation_fraction=0.1, verbose=0, warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 124
        }
      ]
    },
    {
      "metadata": {
        "id": "wMMVrCYCt-RH",
        "colab_type": "code",
        "outputId": "971b3647-2065-4131-9ec2-70888f163d8e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "y_pred_gbr = gbr.predict(X_valid)\n",
        "y_test_gbr = gbr.predict(xtest)\n",
        "print(\"rmsle for gbr: {} \".format(rmsle(y_valid, y_pred_gbr)))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "rmsle for gbr: 0.06530280025903959 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "6s2Lj-HIf8aa",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "df_train['gbr'] = y_pred_gbr\n",
        "df_test['gbr'] = y_test_gbr"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "0RC9puw2gMPW",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "rfr = RandomForestRegressor(n_estimators=30)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "vVD9flBYgZqt",
        "colab_type": "code",
        "outputId": "99bf065e-f953-47b3-9b11-363da56c2491",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 125
        }
      },
      "cell_type": "code",
      "source": [
        "rfr.fit(X_train, y_train)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=None,\n",
              "           max_features='auto', max_leaf_nodes=None,\n",
              "           min_impurity_decrease=0.0, min_impurity_split=None,\n",
              "           min_samples_leaf=1, min_samples_split=2,\n",
              "           min_weight_fraction_leaf=0.0, n_estimators=30, n_jobs=None,\n",
              "           oob_score=False, random_state=None, verbose=0, warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 130
        }
      ]
    },
    {
      "metadata": {
        "id": "32Zrub7mgdjz",
        "colab_type": "code",
        "outputId": "c7f82319-a7d4-4b2a-8a17-86d1dcd9696d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "y_pred_rfr = rfr.predict(X_valid)\n",
        "y_test_rfr = rfr.predict(xtest)\n",
        "print(\"rmsle for gbr: {} \".format(rmsle(y_valid, y_pred_rfr)))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "rmsle for gbr: 0.06035590584753727 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "TntHs_JZglfS",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "df_train['rfr'] = y_pred_rfr\n",
        "df_test['rfr'] = y_test_rfr"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "_MA9YpWKKTop",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Using LGBM"
      ]
    },
    {
      "metadata": {
        "id": "4AsF4X2IKByj",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Configuration of param LGBM"
      ]
    },
    {
      "metadata": {
        "id": "wP3cwJVOI4n8",
        "colab_type": "code",
        "outputId": "7627c97b-96f7-4567-b191-c5744b1e94ee",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "## lightgbm parameters\n",
        "params = {\n",
        "    'learning_rate': 0.1,\n",
        "    'max_depth': 10,\n",
        "    'num_leaves': 60, \n",
        "    'objective': 'regression',\n",
        "    'metric': 'rmse',\n",
        "    'feature_fraction': 0.9,\n",
        "    'bagging_fraction': 0.5,\n",
        "    #'bagging_freq': 5,\n",
        "    'max_bin': 200\n",
        "}\n",
        "\n",
        "## number of rounds\n",
        "nrounds = 1000\n",
        "\n",
        "## train model\n",
        "print('Train model lightGBM...')\n",
        "dtrain = lgb.Dataset(X_train, y_train)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train model lightGBM...\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "2XJlYXl7KN5a",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Traing model LGBM"
      ]
    },
    {
      "metadata": {
        "id": "kC9agpT6oaSD",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "lgbm = lgb.train(params,\n",
        "                dtrain,\n",
        "                num_boost_round = nrounds)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "9XEvil8tKe5w",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Predicting"
      ]
    },
    {
      "metadata": {
        "id": "_sUBHU4yKdaO",
        "colab_type": "code",
        "outputId": "ab66d02e-bb22-48be-fa6a-1068cf1ed2ce",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "## predict valid\n",
        "y_pred_lgbm = lgbm.predict(X_valid)\n",
        "y_test_lgbm = lgbm.predict(xtest)\n",
        "print(\"rmsle for gbr: {} \".format(rmsle(y_valid, y_pred_lgbm)))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "rmsle for gbr: 0.05787703747272962 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "IBFc_W4Cl1gZ",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "df_train['lgbm'] = y_pred_lgbm\n",
        "df_test['lgbm'] = y_test_lgbm"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "r4Txk4mzlsd1",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "## test predictions\n",
        "pred_test = np.exp(lgbm.predict(xtest))\n",
        "\n",
        "## create submission\n",
        "df = pd.DataFrame({'id': id_test, 'trip_duration': pred_test}) \n",
        "df = df.set_index('id')\n",
        "df.to_csv('sub_lgbm.csv', index = True)\n",
        "del(df,pred_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "nzObZsxDNZFX",
        "colab_type": "code",
        "outputId": "04d61e93-f560-4602-9921-0fb4c0d196eb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "cell_type": "code",
      "source": [
        "# submit file\n",
        "!kaggle competitions submit -c nyc-taxi-trip-duration -f sub_lgbm.csv -m \"submission by LGBM version box cox transform\""
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100% 16.9M/16.9M [00:02<00:00, 7.63MB/s]\n",
            "Successfully submitted to New York City Taxi Trip Duration"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "ygXf_Vm3I4q2",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## XGBoost"
      ]
    },
    {
      "metadata": {
        "id": "m-85DoBcK3Nt",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Configuration param of XGB"
      ]
    },
    {
      "metadata": {
        "id": "Vz-vMPqMK9zb",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "## xgb parameters\n",
        "params_xgb = {\n",
        "    'booster':            'gbtree',\n",
        "    'objective':          'reg:linear',\n",
        "    'learning_rate':      0.01,\n",
        "    'max_depth':          14,\n",
        "    'subsample':          0.8,\n",
        "    'colsample_bytree':   0.7,\n",
        "    'colsample_bylevel':  0.7,\n",
        "    'silent':             1\n",
        "}\n",
        "\n",
        "## number of rounds\n",
        "nrounds = 300"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "22DOcjGgLEHh",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Training model"
      ]
    },
    {
      "metadata": {
        "id": "OdxEw-8QLH7w",
        "colab_type": "code",
        "outputId": "b2000968-6cdb-4c98-c037-26174c8ac76c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "## train model\n",
        "print('Train model XGBOOST...')\n",
        "dtrain = xgb.DMatrix(xtrain, y)\n",
        "xgbm = xgb.train(params_xgb,\n",
        "                dtrain,\n",
        "                num_boost_round = nrounds)\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train model XGBOOST...\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "XNH3IHkPLJv1",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Predicting"
      ]
    },
    {
      "metadata": {
        "id": "EcIL4xmPLLi-",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "## test predictions\n",
        "pred_test = np.exp(xgbm.predict(xgb.DMatrix(xtest)))\n",
        "\n",
        "## create submission\n",
        "df = pd.DataFrame({'id': id_test, 'trip_duration': pred_test}) \n",
        "df = df.set_index('id')\n",
        "df.to_csv('sub_xgbm.csv', index = True)\n",
        "del(df,pred_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "LJYBHgXjWLJ3",
        "colab_type": "code",
        "outputId": "ee38121f-f6f8-4a93-e587-09080d3c6094",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "cell_type": "code",
      "source": [
        "# submit file\n",
        "!kaggle competitions submit -c nyc-taxi-trip-duration -f sub_xgbm.csv -m \"submission by XGB model version cox box transform\""
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100% 11.7M/11.7M [00:02<00:00, 5.78MB/s]\n",
            "Successfully submitted to New York City Taxi Trip Duration"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "zMDWfyCgWZEs",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Stacked model (LGBM + XGB)"
      ]
    },
    {
      "metadata": {
        "id": "__wNJW_0gkDo",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "df_lgbm = pd.read_csv(\"sub_lgbm.csv\")\n",
        "df_xgbm = pd.read_csv(\"sub_xgbm.csv\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "GxqMwUNEfbJz",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "lgbm_weight = 0.75\n",
        "xgb_weight = 0.25\n",
        "pred_test = lgbm_weight*df_lgbm['trip_duration'].values + xgb_weight*df_xgbm['trip_duration'].values\n",
        "## create submission\n",
        "df = pd.DataFrame({'id': id_test, 'trip_duration': pred_test}) \n",
        "df = df.set_index('id')\n",
        "df.to_csv('sub_lgbm_xgbm.csv', index = True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "UOPKy1wKgnZB",
        "colab_type": "code",
        "outputId": "4aa8e8e5-b742-40f3-daf7-0ed02367a1f7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "cell_type": "code",
      "source": [
        "# submit file\n",
        "!kaggle competitions submit -c nyc-taxi-trip-duration -f sub_lgbm_xgbm.csv -m \"submission by stacked model LGBM + XGBM\""
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100% 16.9M/16.9M [00:03<00:00, 4.45MB/s]\n",
            "Successfully submitted to New York City Taxi Trip Duration"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "6cnBdgrXjZ8n",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Jf5ZpnOogvs9",
        "colab_type": "code",
        "outputId": "10784667-e103-4e15-ccc0-8a92eb3f8f1d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        }
      },
      "cell_type": "code",
      "source": [
        "print(\"LGBM: Private leadboard {}, Public leadboard {}\".format(0.38622, 0.38879))\n",
        "print(\"XGBM: Private leadboard {}, Public leadboard {}\".format(0.39182, 0.39473))\n",
        "print(\"LGBM + XGBM: Private leadboard {0}, Public leadboard {1}\".format(0.38719, 0.38960))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "LGBM: Private leadboard 0.38622, Public leadboard 0.38879\n",
            "XGBM: Private leadboard 0.39182, Public leadboard 0.39473\n",
            "LGBM + XGBM: Private leadboard 0.38719, Public leadboard 0.3896\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "LErUH0A7jkMt",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "yp5zGNJ-apgy",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Stacked model"
      ]
    },
    {
      "metadata": {
        "id": "MOCeGHwJmMZJ",
        "colab_type": "code",
        "outputId": "c4466f92-eb19-4fe5-ec81-8256dd5235e2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "cell_type": "code",
      "source": [
        "df_train.head()"
      ],
      "execution_count": 0,
      "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>dtr</th>\n",
              "      <th>etr</th>\n",
              "      <th>gbr</th>\n",
              "      <th>rfr</th>\n",
              "      <th>lgbm</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>6.538733</td>\n",
              "      <td>6.284462</td>\n",
              "      <td>6.386696</td>\n",
              "      <td>6.667216</td>\n",
              "      <td>6.630865</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>6.538733</td>\n",
              "      <td>6.284462</td>\n",
              "      <td>6.481406</td>\n",
              "      <td>6.497022</td>\n",
              "      <td>6.564475</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>6.538733</td>\n",
              "      <td>6.284462</td>\n",
              "      <td>6.163007</td>\n",
              "      <td>5.965546</td>\n",
              "      <td>6.134605</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>6.538733</td>\n",
              "      <td>6.427059</td>\n",
              "      <td>6.866480</td>\n",
              "      <td>6.969598</td>\n",
              "      <td>6.980721</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>7.273609</td>\n",
              "      <td>6.427059</td>\n",
              "      <td>8.291095</td>\n",
              "      <td>8.437729</td>\n",
              "      <td>8.453834</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        dtr       etr       gbr       rfr      lgbm\n",
              "0  6.538733  6.284462  6.386696  6.667216  6.630865\n",
              "1  6.538733  6.284462  6.481406  6.497022  6.564475\n",
              "2  6.538733  6.284462  6.163007  5.965546  6.134605\n",
              "3  6.538733  6.427059  6.866480  6.969598  6.980721\n",
              "4  7.273609  6.427059  8.291095  8.437729  8.453834"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 139
        }
      ]
    },
    {
      "metadata": {
        "id": "-vkPtHvMmfnl",
        "colab_type": "code",
        "outputId": "3ab95d5a-3087-402a-d17d-5ad3dc3f83d2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "cell_type": "code",
      "source": [
        "df_test.head()"
      ],
      "execution_count": 0,
      "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>dtr</th>\n",
              "      <th>etr</th>\n",
              "      <th>gbr</th>\n",
              "      <th>rfr</th>\n",
              "      <th>lgbm</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>6.538733</td>\n",
              "      <td>6.427059</td>\n",
              "      <td>6.663997</td>\n",
              "      <td>6.609889</td>\n",
              "      <td>6.714643</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>6.538733</td>\n",
              "      <td>6.627771</td>\n",
              "      <td>6.402525</td>\n",
              "      <td>6.413003</td>\n",
              "      <td>6.459588</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>5.748681</td>\n",
              "      <td>6.427059</td>\n",
              "      <td>6.109192</td>\n",
              "      <td>5.997409</td>\n",
              "      <td>6.101380</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>7.273609</td>\n",
              "      <td>6.427059</td>\n",
              "      <td>7.061644</td>\n",
              "      <td>6.915780</td>\n",
              "      <td>6.862266</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5.748681</td>\n",
              "      <td>6.284462</td>\n",
              "      <td>5.882973</td>\n",
              "      <td>5.810722</td>\n",
              "      <td>5.903209</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        dtr       etr       gbr       rfr      lgbm\n",
              "0  6.538733  6.427059  6.663997  6.609889  6.714643\n",
              "1  6.538733  6.627771  6.402525  6.413003  6.459588\n",
              "2  5.748681  6.427059  6.109192  5.997409  6.101380\n",
              "3  7.273609  6.427059  7.061644  6.915780  6.862266\n",
              "4  5.748681  6.284462  5.882973  5.810722  5.903209"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 140
        }
      ]
    },
    {
      "metadata": {
        "id": "r-q9TcyvmiP4",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "en = ElasticNet(alpha=0.2)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "K5n680crm5zW",
        "colab_type": "code",
        "outputId": "68524809-c440-4672-af65-b19c6ae4c32d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        }
      },
      "cell_type": "code",
      "source": [
        "en.fit(df_train,y_valid)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "ElasticNet(alpha=0.2, copy_X=True, fit_intercept=True, l1_ratio=0.5,\n",
              "      max_iter=1000, normalize=False, positive=False, precompute=False,\n",
              "      random_state=None, selection='cyclic', tol=0.0001, warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 150
        }
      ]
    },
    {
      "metadata": {
        "id": "iVMLM8ylnC3y",
        "colab_type": "code",
        "outputId": "bcd1f4a3-0fc8-437b-b3cd-c7cab655c9f4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "en_pred = en.predict(df_train)\n",
        "print(\"rmsle for gbr: {} \".format(rmsle(y_valid, en_pred)))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "rmsle for gbr: 0.06528193971999488 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "AvehbhGdnPDp",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "pred_test = np.exp(en.predict(df_test))\n",
        "## create submission\n",
        "df = pd.DataFrame({'id': id_test, 'trip_duration': pred_test}) \n",
        "df = df.set_index('id')\n",
        "df.to_csv('baseline_model_stacked.csv', index = True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "ivZbjU_wnnl6",
        "colab_type": "code",
        "outputId": "335c45ba-ec7c-4373-de71-2bee18b33dbd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "cell_type": "code",
      "source": [
        "# submit file\n",
        "!kaggle competitions submit -c nyc-taxi-trip-duration -f baseline_model_stacked.csv -m \"submission by stacked model\""
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100% 16.8M/16.8M [00:03<00:00, 5.64MB/s]\n",
            "Successfully submitted to New York City Taxi Trip Duration"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "NYgfv1o9ot8K",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}