Shop_Floor_v1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.4"
    },
    "colab": {
      "name": "Shop_Floor-v1.ipynb",
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/mohan-mj/Manufacturing-Line-I4.0/blob/master/Shop_Floor_v1.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oOhHcon8Pm_3",
        "colab_type": "text"
      },
      "source": [
        "Shop floor always have challenges. When the part fails at the end of line testing, shop floor manager instinctively retrace your steps to identify at what point you went wrong. Here, the ML based algorithm is closely monitoring the parts as they progress through the manufacturing processes. So that the defective parts can be intelligently identified."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "akJ8fvsfGVdn",
        "colab_type": "code",
        "outputId": "ec38f43d-1a6a-42b7-8d25-14e4d87f04bd",
        "colab": {
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/files.js": {
              "data": "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",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "status": 200,
              "status_text": ""
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 113
        }
      },
      "source": [
        "from google.colab import files\n",
        "files.upload()"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-b8788903-6289-4742-9d7a-588bd8110e4c\" name=\"files[]\" multiple disabled />\n",
              "     <output id=\"result-b8788903-6289-4742-9d7a-588bd8110e4c\">\n",
              "      Upload widget is only available when the cell has been executed in the\n",
              "      current browser session. Please rerun this cell to enable.\n",
              "      </output>\n",
              "      <script src=\"/nbextensions/google.colab/files.js\"></script> "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Saving kaggle.json to kaggle.json\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'kaggle.json': b'{\"username\":\"jidhumohan\",\"key\":\"a034f244ffcb02806368a9a420a631f4\"}'}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cDLXrqcNGlU7",
        "colab_type": "code",
        "outputId": "eaa9496d-f75d-4f3e-cbc4-b2e9427e5d53",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "# Let's make sure the kaggle.json file is present.\n",
        "!ls -lha kaggle.json"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "-rw-r--r-- 1 root root 66 Sep 18 08:08 kaggle.json\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QITWh1N5GoqG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Next, install the Kaggle API client.\n",
        "!pip install -q kaggle"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KixiXyagGy7Y",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# The Kaggle API client expects this file to be in ~/.kaggle,\n",
        "# so move it there.\n",
        "!mkdir -p ~/.kaggle\n",
        "!cp kaggle.json ~/.kaggle/\n",
        "\n",
        "# This permissions change avoids a warning on Kaggle tool startup.\n",
        "!chmod 600 ~/.kaggle/kaggle.json"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fREG6xVYZRmE",
        "colab_type": "code",
        "outputId": "434bac37-2cbe-4838-b35c-91d782edb6c3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 381
        }
      },
      "source": [
        "!kaggle competitions download -c bosch-production-line-performance"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading test_categorical.csv.zip to /content\n",
            " 50% 9.00M/18.1M [00:00<00:00, 14.9MB/s]\n",
            "100% 18.1M/18.1M [00:00<00:00, 26.4MB/s]\n",
            "Downloading train_categorical.csv.zip to /content\n",
            " 50% 9.00M/18.1M [00:00<00:00, 13.4MB/s]\n",
            "100% 18.1M/18.1M [00:00<00:00, 23.1MB/s]\n",
            "Downloading train_date.csv.zip to /content\n",
            " 77% 45.0M/58.8M [00:01<00:00, 21.9MB/s]\n",
            "100% 58.8M/58.8M [00:01<00:00, 41.9MB/s]\n",
            "Downloading test_date.csv.zip to /content\n",
            " 83% 49.0M/58.8M [00:02<00:00, 17.3MB/s]\n",
            "100% 58.8M/58.8M [00:02<00:00, 27.7MB/s]\n",
            "Downloading sample_submission.csv.zip to /content\n",
            "  0% 0.00/1.55M [00:00<?, ?B/s]\n",
            "100% 1.55M/1.55M [00:00<00:00, 51.1MB/s]\n",
            "Downloading test_numeric.csv.zip to /content\n",
            " 97% 261M/270M [00:05<00:00, 40.1MB/s]\n",
            "100% 270M/270M [00:05<00:00, 47.4MB/s]\n",
            "Downloading train_numeric.csv.zip to /content\n",
            " 98% 265M/270M [00:05<00:00, 38.9MB/s]\n",
            "100% 270M/270M [00:05<00:00, 47.7MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HfI6qs96bPvX",
        "colab_type": "text"
      },
      "source": [
        "### Read Data - 1Million data points"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e3Zwtc8byC6a",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Import Libs\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "from xgboost import XGBClassifier\n",
        "from sklearn.metrics import matthews_corrcoef, roc_auc_score\n",
        "from sklearn.model_selection import cross_val_score, StratifiedKFold\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import zipfile\n",
        "%matplotlib inline"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i-XjvXEcMpnu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# #In case data is from My Drive\n",
        "# from google.colab import drive\n",
        "# drive.mount('/content/drive')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jEAVdJi6MlKQ",
        "colab_type": "code",
        "outputId": "cb5edd02-bc69-4323-da46-5bac95a41713",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        }
      },
      "source": [
        "!ls"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "kaggle.json\t\t   test_categorical.csv.zip  train_categorical.csv.zip\n",
            "sample_data\t\t   test_date.csv.zip\t     train_date.csv.zip\n",
            "sample_submission.csv.zip  test_numeric.csv.zip      train_numeric.csv.zip\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L3uubgecyC6e",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "zf1 = zipfile.ZipFile('train_date.csv.zip') \n",
        "date_chunks = pd.read_csv(zf1.open('train_date.csv'),\n",
        "                          index_col=0, chunksize=100000, dtype=np.float32)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZJNeoHYVyC6h",
        "colab_type": "code",
        "outputId": "665abca6-267c-410c-cad6-3629c23d66f7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 173
        }
      },
      "source": [
        "for chunk in date_chunks:\n",
        "    #print(len(chunk))\n",
        "    print(chunk.head())\n",
        "    break"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "         L0_S0_D1     L0_S0_D3  ...  L3_S51_D4261  L3_S51_D4263\n",
            "Id                              ...                            \n",
            "4.0     82.239998    82.239998  ...           NaN           NaN\n",
            "6.0           NaN          NaN  ...           NaN           NaN\n",
            "7.0   1618.699951  1618.699951  ...           NaN           NaN\n",
            "9.0   1149.199951  1149.199951  ...           NaN           NaN\n",
            "11.0   602.640015   602.640015  ...           NaN           NaN\n",
            "\n",
            "[5 rows x 1156 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CiKCHxEANhM1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# s = 0\n",
        "# for chunk in date_chunks:\n",
        "#     s += len(chunk)\n",
        "# print(\"Total data size : \", s)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JGj5EP4xyC6k",
        "colab_type": "text"
      },
      "source": [
        "The data represents measurements of parts as they move through production lines. Each part has a unique Id. The goal is to predict which parts will fail quality control (represented by a 'Response' = 1).\n",
        "\n",
        "The dataset contains an extremely large number of anonymized features. Features are named according to a convention that tells you the production line, the station on the line, and a feature number. E.g. L3_S36_F3939 is a feature measured on line 3, station 36, and is feature number 3939.\n",
        "\n",
        "On account of the large size of the dataset, we have separated the files by the type of feature they contain: numerical, categorical, and finally, a file with date features. The date features provide a timestamp for when each measurement was taken. Each date column ends in a number that corresponds to the previous feature number. E.g. the value of L0_S0_D1 is the time at which L0_S0_F0 was taken."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CJ0gfkPayC6l",
        "colab_type": "text"
      },
      "source": [
        "**Description**\n",
        "\n",
        "train_numeric.csv - the training set numeric features (this file contains the 'Response' variable)\n",
        "\n",
        "test_numeric.csv - the test set numeric features (predict the 'Response' for these Ids)\n",
        "\n",
        "train_categorical.csv - the training set categorical features\n",
        "\n",
        "test_categorical.csv - the test set categorical features\n",
        "\n",
        "train_date.csv - the training set date features\n",
        "\n",
        "test_date.csv - the test set date features\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dnCdT5mQyC6m",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "zf2 = zipfile.ZipFile('train_numeric.csv.zip') \n",
        "num_chunks = pd.read_csv(zf2.open('train_numeric.csv'),\n",
        "                          index_col=0, usecols=list(range(969)), chunksize=100000, dtype=np.float32)\n",
        "#num_chunks = pd.read_csv(\"train_numeric.csv\", index_col=0, usecols=list(range(969)), chunksize=100000, dtype=np.float32)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1X25tMZCyC6o",
        "colab_type": "code",
        "outputId": "d886e710-3873-4154-a46b-3b8aeae2f3da",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 211
        }
      },
      "source": [
        "for chunk in num_chunks:\n",
        "    print(print(chunk.head()))\n",
        "    break"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "      L0_S0_F0  L0_S0_F2  L0_S0_F4  ...  L3_S51_F4258  L3_S51_F4260  L3_S51_F4262\n",
            "Id                                  ...                                          \n",
            "4.0      0.030    -0.034    -0.197  ...           NaN           NaN           NaN\n",
            "6.0        NaN       NaN       NaN  ...           NaN           NaN           NaN\n",
            "7.0      0.088     0.086     0.003  ...           NaN           NaN           NaN\n",
            "9.0     -0.036    -0.064     0.294  ...           NaN           NaN           NaN\n",
            "11.0    -0.055    -0.086     0.294  ...           NaN           NaN           NaN\n",
            "\n",
            "[5 rows x 968 columns]\n",
            "None\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DKnTdzUmyC6s",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X = pd.concat([pd.concat([dchunk, nchunk], axis=1).sample(frac=0.1) for dchunk, nchunk in zip(date_chunks, num_chunks)])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1DnGGEfUyC6u",
        "colab_type": "code",
        "outputId": "26c206d5-6f33-46c7-d321-8b34cb91cd33",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        }
      },
      "source": [
        "#zf = zipfile.ZipFile('train_numeric.csv.zip') \n",
        "y = pd.read_csv(zf2.open('train_numeric.csv'), index_col=0, usecols=[0,969], dtype=np.float32).loc[X.index].values.ravel()\n",
        "# y = pd.read_csv(\"train_numeric.csv\", index_col=0, usecols=[0,969], dtype=np.float32).loc[X.index].values.ravel()"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/numpy/lib/arraysetops.py:569: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
            "  mask |= (ar1 == a)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9-iDtD7wyC6w",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "bbb81775-692f-4e31-bb79-736a62423a0e"
      },
      "source": [
        "X = X.values\n",
        "y"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0., 0., 0., ..., 0., 0., 0.], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TmSF_QdLyC62",
        "colab_type": "code",
        "outputId": "073dd159-33a0-451d-891c-d04b2ca29145",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "y.sum()"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "639.0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "21D-itHJyC64",
        "colab_type": "code",
        "outputId": "6a0a9de2-9f52-4f94-c52c-7e08a631114b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "len(y)"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "108375"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VvCOEBk-yC68",
        "colab_type": "code",
        "outputId": "19296487-5752-487f-efd7-4e26eb46662d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "print('We have {}% defective parts (in 0.1 fraction of random sample)'.format(y.sum()*100/len(y)))"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "We have 0.5896193771626298% defective parts (in 0.1 fraction of random sample)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X6VWo9dNyC6-",
        "colab_type": "code",
        "outputId": "cd784d48-9332-443b-9401-5d5a15f710a2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 139
        }
      },
      "source": [
        "clf = XGBClassifier(base_score=0.005)\n",
        "clf.fit(X, y)"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "XGBClassifier(base_score=0.005, booster='gbtree', colsample_bylevel=1,\n",
              "              colsample_bynode=1, colsample_bytree=1, gamma=0,\n",
              "              learning_rate=0.1, max_delta_step=0, max_depth=3,\n",
              "              min_child_weight=1, missing=None, n_estimators=100, n_jobs=1,\n",
              "              nthread=None, objective='binary:logistic', random_state=0,\n",
              "              reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
              "              silent=None, subsample=1, verbosity=1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ibepTehcyC7A",
        "colab_type": "text"
      },
      "source": [
        "### Save Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_vYRQrJTyC7C",
        "colab_type": "code",
        "outputId": "fea10d0b-ec44-4860-940e-4360640f6397",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "import sys\n",
        "print(sys.version)"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "3.6.8 (default, Jan 14 2019, 11:02:34) \n",
            "[GCC 8.0.1 20180414 (experimental) [trunk revision 259383]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hy6cZXn0yC7G",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pickle"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4-YCxP-QyC7J",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# save the model to disk\n",
        "filename = 'XGBoost_Model.sav'\n",
        "pickle.dump(clf, open(filename, 'wb'))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tsMbRcS6yC7P",
        "colab_type": "text"
      },
      "source": [
        "### Load Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "58ju7I-8yC7R",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# load the model from disk\n",
        "myModel = pickle.load(open(filename, 'rb'))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MSm0Si-PyC7S",
        "colab_type": "code",
        "outputId": "91fa9569-fd1a-4b1a-bcc6-41f1347821df",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 339
        }
      },
      "source": [
        "# threshold for a manageable number of features\n",
        "plt.hist(myModel.feature_importances_[myModel.feature_importances_>0])\n",
        "important_indices = np.where(myModel.feature_importances_>0.005)[0]\n",
        "print(important_indices)"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[  36  205  402  883 1018 1019 1042 1050 1056 1165 1168 1173 1189 1190\n",
            " 1191 1215 1226 1294 1312 1441 1477 1488 1511 1512 1516 1518 1539 1548\n",
            " 1549 1550 1849 1856 1858 1879 1881 1889 1891 1911 1921 1927 1932 1942\n",
            " 1947 1949 1962 1974 1975 1985 1992 1993 1995 2006 2010 2039]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD8CAYAAAB5Pm/hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nutlMiaXuAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LxNJVT01s-PV",
        "colab_type": "code",
        "outputId": "05116641-5eb3-43f5-d5ec-36510dbbd6e5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        }
      },
      "source": [
        "np.where(myModel.feature_importances_>0.005)"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([  36,  205,  402,  883, 1018, 1019, 1042, 1050, 1056, 1165, 1168,\n",
              "        1173, 1189, 1190, 1191, 1215, 1226, 1294, 1312, 1441, 1477, 1488,\n",
              "        1511, 1512, 1516, 1518, 1539, 1548, 1549, 1550, 1849, 1856, 1858,\n",
              "        1879, 1881, 1889, 1891, 1911, 1921, 1927, 1932, 1942, 1947, 1949,\n",
              "        1962, 1974, 1975, 1985, 1992, 1993, 1995, 2006, 2010, 2039]),)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VBsB0mv4yC7V",
        "colab_type": "text"
      },
      "source": [
        "### Model for entire data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5l-_TccqyC7W",
        "colab_type": "code",
        "outputId": "28eab8a8-13ca-429d-a0d2-1d31562cca9e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        }
      },
      "source": [
        "# load entire dataset for these features. \n",
        "# note where the feature indices are split so we can load the correct ones straight from read_csv\n",
        "n_date_features = 1156\n",
        "X_full = np.concatenate([\n",
        "    pd.read_csv(zf1.open(\"train_date.csv\"), index_col=0, dtype=np.float32,\n",
        "                usecols=np.concatenate([[0], important_indices[important_indices < n_date_features] + 1])).values,\n",
        "    pd.read_csv(zf2.open(\"train_numeric.csv\"), index_col=0, dtype=np.float32,\n",
        "                usecols=np.concatenate([[0], important_indices[important_indices >= n_date_features] + 1 - 1156])).values\n",
        "                ], axis=1)\n",
        "y_full = pd.read_csv(zf2.open(\"train_numeric.csv\"), index_col=0, dtype=np.float32, usecols=[0,969]).values.ravel()"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/numpy/lib/arraysetops.py:569: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
            "  mask |= (ar1 == a)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TJWPBRIlyC7Z",
        "colab_type": "code",
        "outputId": "6806cd1e-7a3c-4713-d360-5cfb7638582b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 176
        }
      },
      "source": [
        "clf_full_data = XGBClassifier(max_depth=5, base_score=0.005)\n",
        "skf = StratifiedKFold()\n",
        "skf.get_n_splits(X_full, y_full)\n",
        "preds = np.ones(y_full.shape[0])\n",
        "\n",
        "for train_index, test_index in skf.split(X_full, y_full):\n",
        "    print(\"TRAIN:\", train_index, \"TEST:\", test_index)\n",
        "    #X_train, X_test = X[train_index], X[test_index]\n",
        "    #y_train, y_test = y[train_index], y[test_index]\n",
        "    \n",
        "    preds[test_index] = clf_full_data.fit(X_full[train_index], y_full[train_index]).predict_proba(X_full[test_index])[:,1]\n",
        "    print(\"ROC AUC: \", roc_auc_score(y_full[test_index], preds[test_index]))"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/model_selection/_split.py:629: FutureWarning: The default value of n_split will change from 3 to 5 in version 0.22. Specify it explicitly to silence this warning.\n",
            "  warnings.warn(NSPLIT_WARNING, FutureWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "TRAIN: [ 394571  394572  394573 ... 1183744 1183745 1183746] TEST: [     0      1      2 ... 395664 395665 395695]\n",
            "ROC AUC:  0.7258777863733405\n",
            "TRAIN: [      0       1       2 ... 1183744 1183745 1183746] TEST: [394571 394572 394573 ... 790831 790909 790942]\n",
            "ROC AUC:  0.7047022699007283\n",
            "TRAIN: [     0      1      2 ... 790831 790909 790942] TEST: [ 789150  789151  789152 ... 1183744 1183745 1183746]\n",
            "ROC AUC:  0.69761564050326\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rFxW0yVMyC7h",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "85c9d33b-1360-4300-a059-a0aed7b829f1"
      },
      "source": [
        "print(roc_auc_score(y_full, preds))"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.7090342529625953\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QBRUmpvSyC7m",
        "colab_type": "text"
      },
      "source": [
        "### Save Model - Full Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mI5w-EfkyC7m",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# save the model to disk\n",
        "filename = 'XGBoost_Model_Full_Data.sav'\n",
        "pickle.dump(clf_full_data, open(filename, 'wb'))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xeS-R2OxyC7r",
        "colab_type": "text"
      },
      "source": [
        "### Load Model - Full Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8FXD9NbIyC7s",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# load the model from disk\n",
        "clf_full_data = pickle.load(open(filename, 'rb'))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DH_oMgH7pvfg",
        "colab_type": "text"
      },
      "source": [
        "### Best Threshold Selection"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DMEw0n0ayC7w",
        "colab_type": "code",
        "outputId": "2df6bbcf-c135-4a05-f064-573cfc3ec945",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 287
        }
      },
      "source": [
        "# pick the best threshold out-of-fold\n",
        "thresholds = np.linspace(0.01, 0.99, 50)\n",
        "mcc = np.array([matthews_corrcoef(y_full, preds>thr) for thr in thresholds])\n",
        "plt.plot(thresholds, mcc)\n",
        "best_threshold = thresholds[mcc.argmax()]\n",
        "print(mcc.max())"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.20415193242539573\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAD8CAYAAAB3u9PLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAWCiIgEFAgiIgLA/wfxPKuP89Yx8gAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vWUrKf_xyC7z",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import confusion_matrix, roc_curve, auc"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LbLpJZ5-yC74",
        "colab_type": "code",
        "outputId": "a0588395-7c3a-4825-eec5-4acc45685ba4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "confusion_matrix(y_full, preds>0.08)"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[1172705,    4163],\n",
              "       [   5613,    1266]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z8nHndFOyC76",
        "colab_type": "code",
        "outputId": "a5112b22-babd-4c69-b481-68e1d7594bda",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "y_full.shape"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1183747,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YzjLhUt_yC78",
        "colab_type": "code",
        "outputId": "6e5424c0-42f4-4636-b1d4-e662fe38b3e7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "y_full.sum()"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "6879.0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0Byoc0s0yC7_",
        "colab_type": "code",
        "outputId": "dbae1123-bc2f-412a-ea20-490fc8e6494b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "preds[preds>.02].shape"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(9301,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ymAPb8m3yC8C",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "fpr, tpr, threshold = roc_curve(y_full, preds)\n",
        "roc_auc = auc(fpr, tpr)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3xFtwb73yC8D",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%config InlineBackend.figure_format = 'retina'"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "86jcgWdsyC8F",
        "colab_type": "code",
        "outputId": "88f25558-537c-499f-e455-b004af267b5d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 294
        }
      },
      "source": [
        "plt.figure()\n",
        "plt.plot(fpr, tpr, color='darkorange',\n",
        "         lw=2, label='ROC curve (area = %0.2f)' % roc_auc)\n",
        "plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--')\n",
        "plt.plot(fpr, threshold, color='red', lw=2, linestyle='--')\n",
        "plt.xlim([0.0, 1.0])\n",
        "plt.ylim([0.0, 1.05])\n",
        "plt.xlabel('False Positive Rate')\n",
        "plt.ylabel('True Positive Rate')\n",
        "plt.title('Receiver operating characteristic (ROC)')\n",
        "plt.legend(loc=\"lower right\")\n",
        "plt.show()"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxMAAAIqCAYAAABfWBeBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 393,
              "height": 277
            }
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JDh62l9uyC8H",
        "colab_type": "code",
        "outputId": "18280833-0613-4d91-f167-85eb7cd058cb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 287
        }
      },
      "source": [
        "plt.plot(fpr, threshold, color='red', lw=2, linestyle='--')\n",
        "plt.ylim([0.0, 0.02])"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.0, 0.02)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 42
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxIAAAH4CAYAAADehNsGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\npGQKJEREREREpGT/H7WOiof7PCxBAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 393,
              "height": 252
            }
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UcdGWoPiyC8K",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}