conv1d-generic-houses-prices.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:11.463213Z",
     "start_time": "2020-09-19T17:12:11.378123Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 1;\n",
       "                var nbb_unformatted_code = \"%load_ext nb_black\\n%load_ext autoreload\\n\\n%autoreload 2\";\n",
       "                var nbb_formatted_code = \"%load_ext nb_black\\n%load_ext autoreload\\n\\n%autoreload 2\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%load_ext nb_black\n",
    "%load_ext autoreload\n",
    "\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:12.694106Z",
     "start_time": "2020-09-19T17:12:11.464581Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 2;\n",
       "                var nbb_unformatted_code = \"import os\\nfrom pathlib import Path\\n\\nfrom requests import get\\nimport pandas as pd\\nimport numpy as np\\n\\nfrom sklearn.metrics import mean_absolute_error, mean_squared_error\\nfrom sklearn.preprocessing import LabelEncoder, StandardScaler\\nfrom tensorflow.keras.utils import plot_model\\nfrom tensorflow.keras.callbacks import EarlyStopping\\n\\nimport logging\\n\\nlogging.basicConfig(level=logging.WARN)\";\n",
       "                var nbb_formatted_code = \"import os\\nfrom pathlib import Path\\n\\nfrom requests import get\\nimport pandas as pd\\nimport numpy as np\\n\\nfrom sklearn.metrics import mean_absolute_error, mean_squared_error\\nfrom sklearn.preprocessing import LabelEncoder, StandardScaler\\nfrom tensorflow.keras.utils import plot_model\\nfrom tensorflow.keras.callbacks import EarlyStopping\\n\\nimport logging\\n\\nlogging.basicConfig(level=logging.WARN)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import os\n",
    "from pathlib import Path\n",
    "\n",
    "from requests import get\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from sklearn.metrics import mean_absolute_error, mean_squared_error\n",
    "from sklearn.preprocessing import LabelEncoder, StandardScaler\n",
    "from tensorflow.keras.utils import plot_model\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "\n",
    "import logging\n",
    "\n",
    "logging.basicConfig(level=logging.WARN)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:12.879205Z",
     "start_time": "2020-09-19T17:12:12.695765Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 3;\n",
       "                var nbb_unformatted_code = \"from xplainet.input_utils import preproc_dataset\\nfrom xplainet.model import build_model\\nfrom xplainet.random_utils import setup_seed, SEED\\nfrom sklearn.model_selection import ShuffleSplit\\n\\nimport matplotlib.pyplot as plt\\nfrom matplotlib.pyplot import imshow\\n\\n%matplotlib inline\\n\\nfrom zipfile import ZipFile\";\n",
       "                var nbb_formatted_code = \"from xplainet.input_utils import preproc_dataset\\nfrom xplainet.model import build_model\\nfrom xplainet.random_utils import setup_seed, SEED\\nfrom sklearn.model_selection import ShuffleSplit\\n\\nimport matplotlib.pyplot as plt\\nfrom matplotlib.pyplot import imshow\\n\\n%matplotlib inline\\n\\nfrom zipfile import ZipFile\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from xplainet.input_utils import preproc_dataset\n",
    "from xplainet.model import build_model\n",
    "from xplainet.random_utils import setup_seed, SEED\n",
    "from sklearn.model_selection import ShuffleSplit\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pyplot import imshow\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "from zipfile import ZipFile"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:12.902319Z",
     "start_time": "2020-09-19T17:12:12.880236Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 4;\n",
       "                var nbb_unformatted_code = \"setup_seed()\";\n",
       "                var nbb_formatted_code = \"setup_seed()\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "setup_seed()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:12.986358Z",
     "start_time": "2020-09-19T17:12:12.903233Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 5;\n",
       "                var nbb_unformatted_code = \"def download(url, out, force=False, verify=True):\\n    out.parent.mkdir(parents=True, exist_ok=True)\\n    if force and out.exists():\\n        print(f\\\"Removing file at {str(out)}\\\")\\n        out.unlink()\\n\\n    if out.exists():\\n        print(\\\"File already exists.\\\")\\n        return\\n    print(f\\\"Downloading {url} at {str(out)} ...\\\")\\n    # open in binary mode\\n    with out.open(mode=\\\"wb\\\") as file:\\n        # get request\\n        response = get(url, verify=verify)\\n        for chunk in response.iter_content(100000):\\n            # write to file\\n            file.write(chunk)\";\n",
       "                var nbb_formatted_code = \"def download(url, out, force=False, verify=True):\\n    out.parent.mkdir(parents=True, exist_ok=True)\\n    if force and out.exists():\\n        print(f\\\"Removing file at {str(out)}\\\")\\n        out.unlink()\\n\\n    if out.exists():\\n        print(\\\"File already exists.\\\")\\n        return\\n    print(f\\\"Downloading {url} at {str(out)} ...\\\")\\n    # open in binary mode\\n    with out.open(mode=\\\"wb\\\") as file:\\n        # get request\\n        response = get(url, verify=verify)\\n        for chunk in response.iter_content(100000):\\n            # write to file\\n            file.write(chunk)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def download(url, out, force=False, verify=True):\n",
    "    out.parent.mkdir(parents=True, exist_ok=True)\n",
    "    if force and out.exists():\n",
    "        print(f\"Removing file at {str(out)}\")\n",
    "        out.unlink()\n",
    "\n",
    "    if out.exists():\n",
    "        print(\"File already exists.\")\n",
    "        return\n",
    "    print(f\"Downloading {url} at {str(out)} ...\")\n",
    "    # open in binary mode\n",
    "    with out.open(mode=\"wb\") as file:\n",
    "        # get request\n",
    "        response = get(url, verify=verify)\n",
    "        for chunk in response.iter_content(100000):\n",
    "            # write to file\n",
    "            file.write(chunk)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.031207Z",
     "start_time": "2020-09-19T17:12:12.987309Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 6;\n",
       "                var nbb_unformatted_code = \"def plot_history(history):\\n    loss_list = [s for s in history.history.keys() if \\\"loss\\\" in s and \\\"val\\\" not in s]\\n    val_loss_list = [s for s in history.history.keys() if \\\"loss\\\" in s and \\\"val\\\" in s]\\n    acc_list = [s for s in history.history.keys() if \\\"AUC\\\" in s and \\\"val\\\" not in s]\\n    val_acc_list = [s for s in history.history.keys() if \\\"AUC\\\" in s and \\\"val\\\" in s]\\n\\n    if len(loss_list) == 0:\\n        print(\\\"Loss is missing in history\\\")\\n        return\\n\\n    ## As loss always exists\\n    epochs = range(1, len(history.history[loss_list[0]]) + 1)\\n\\n    ## Loss\\n    plt.figure(1)\\n    for l in loss_list:\\n        plt.plot(\\n            epochs,\\n            history.history[l],\\n            \\\"b\\\",\\n            label=\\\"Training loss (\\\"\\n            + str(str(format(history.history[l][-1], \\\".5f\\\")) + \\\")\\\"),\\n        )\\n    for l in val_loss_list:\\n        plt.plot(\\n            epochs,\\n            history.history[l],\\n            \\\"g\\\",\\n            label=\\\"Validation loss (\\\"\\n            + str(str(format(history.history[l][-1], \\\".5f\\\")) + \\\")\\\"),\\n        )\\n\\n    plt.title(\\\"Loss\\\")\\n    plt.xlabel(\\\"Epochs\\\")\\n    plt.ylabel(\\\"Loss\\\")\\n    plt.legend()\\n\\n    plt.show()\";\n",
       "                var nbb_formatted_code = \"def plot_history(history):\\n    loss_list = [s for s in history.history.keys() if \\\"loss\\\" in s and \\\"val\\\" not in s]\\n    val_loss_list = [s for s in history.history.keys() if \\\"loss\\\" in s and \\\"val\\\" in s]\\n    acc_list = [s for s in history.history.keys() if \\\"AUC\\\" in s and \\\"val\\\" not in s]\\n    val_acc_list = [s for s in history.history.keys() if \\\"AUC\\\" in s and \\\"val\\\" in s]\\n\\n    if len(loss_list) == 0:\\n        print(\\\"Loss is missing in history\\\")\\n        return\\n\\n    ## As loss always exists\\n    epochs = range(1, len(history.history[loss_list[0]]) + 1)\\n\\n    ## Loss\\n    plt.figure(1)\\n    for l in loss_list:\\n        plt.plot(\\n            epochs,\\n            history.history[l],\\n            \\\"b\\\",\\n            label=\\\"Training loss (\\\"\\n            + str(str(format(history.history[l][-1], \\\".5f\\\")) + \\\")\\\"),\\n        )\\n    for l in val_loss_list:\\n        plt.plot(\\n            epochs,\\n            history.history[l],\\n            \\\"g\\\",\\n            label=\\\"Validation loss (\\\"\\n            + str(str(format(history.history[l][-1], \\\".5f\\\")) + \\\")\\\"),\\n        )\\n\\n    plt.title(\\\"Loss\\\")\\n    plt.xlabel(\\\"Epochs\\\")\\n    plt.ylabel(\\\"Loss\\\")\\n    plt.legend()\\n\\n    plt.show()\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_history(history):\n",
    "    loss_list = [s for s in history.history.keys() if \"loss\" in s and \"val\" not in s]\n",
    "    val_loss_list = [s for s in history.history.keys() if \"loss\" in s and \"val\" in s]\n",
    "    acc_list = [s for s in history.history.keys() if \"AUC\" in s and \"val\" not in s]\n",
    "    val_acc_list = [s for s in history.history.keys() if \"AUC\" in s and \"val\" in s]\n",
    "\n",
    "    if len(loss_list) == 0:\n",
    "        print(\"Loss is missing in history\")\n",
    "        return\n",
    "\n",
    "    ## As loss always exists\n",
    "    epochs = range(1, len(history.history[loss_list[0]]) + 1)\n",
    "\n",
    "    ## Loss\n",
    "    plt.figure(1)\n",
    "    for l in loss_list:\n",
    "        plt.plot(\n",
    "            epochs,\n",
    "            history.history[l],\n",
    "            \"b\",\n",
    "            label=\"Training loss (\"\n",
    "            + str(str(format(history.history[l][-1], \".5f\")) + \")\"),\n",
    "        )\n",
    "    for l in val_loss_list:\n",
    "        plt.plot(\n",
    "            epochs,\n",
    "            history.history[l],\n",
    "            \"g\",\n",
    "            label=\"Validation loss (\"\n",
    "            + str(str(format(history.history[l][-1], \".5f\")) + \")\"),\n",
    "        )\n",
    "\n",
    "    plt.title(\"Loss\")\n",
    "    plt.xlabel(\"Epochs\")\n",
    "    plt.ylabel(\"Loss\")\n",
    "    plt.legend()\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Bank marketing : loading data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.072505Z",
     "start_time": "2020-09-19T17:12:13.032172Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 7;\n",
       "                var nbb_unformatted_code = \"dataset_name = \\\"kaggle-house-prices\\\"\\nout = Path(os.getcwd() + \\\"/data/\\\" + dataset_name + \\\"-train.csv\\\")\\n\\ntarget = \\\"SalePrice\\\"\\nto_remove = []\\ntrain = pd.read_csv(out, sep=\\\",\\\", low_memory=False)\";\n",
       "                var nbb_formatted_code = \"dataset_name = \\\"kaggle-house-prices\\\"\\nout = Path(os.getcwd() + \\\"/data/\\\" + dataset_name + \\\"-train.csv\\\")\\n\\ntarget = \\\"SalePrice\\\"\\nto_remove = []\\ntrain = pd.read_csv(out, sep=\\\",\\\", low_memory=False)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataset_name = \"kaggle-house-prices\"\n",
    "out = Path(os.getcwd() + \"/data/\" + dataset_name + \"-train.csv\")\n",
    "\n",
    "target = \"SalePrice\"\n",
    "to_remove = []\n",
    "train = pd.read_csv(out, sep=\",\", low_memory=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.308917Z",
     "start_time": "2020-09-19T17:12:13.074139Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "census-income.csv\tforest-cover-type.zip\t       wine-quality.csv\r\n",
      "census-income_test.csv\tkaggle-house-prices-test.csv\r\n",
      "forest-cover-type.csv\tkaggle-house-prices-train.csv\r\n"
     ]
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 8;\n",
       "                var nbb_unformatted_code = \"!ls data\";\n",
       "                var nbb_formatted_code = \"!ls data\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "!ls data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.364073Z",
     "start_time": "2020-09-19T17:12:13.314701Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1460, 81)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 9;\n",
       "                var nbb_unformatted_code = \"train.shape\";\n",
       "                var nbb_formatted_code = \"train.shape\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.389005Z",
     "start_time": "2020-09-19T17:12:13.365104Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Id', 'MSSubClass', 'MSZoning', 'LotFrontage', 'LotArea', 'Street',\n",
       "       'Alley', 'LotShape', 'LandContour', 'Utilities', 'LotConfig',\n",
       "       'LandSlope', 'Neighborhood', 'Condition1', 'Condition2', 'BldgType',\n",
       "       'HouseStyle', 'OverallQual', 'OverallCond', 'YearBuilt', 'YearRemodAdd',\n",
       "       'RoofStyle', 'RoofMatl', 'Exterior1st', 'Exterior2nd', 'MasVnrType',\n",
       "       'MasVnrArea', 'ExterQual', 'ExterCond', 'Foundation', 'BsmtQual',\n",
       "       'BsmtCond', 'BsmtExposure', 'BsmtFinType1', 'BsmtFinSF1',\n",
       "       'BsmtFinType2', 'BsmtFinSF2', 'BsmtUnfSF', 'TotalBsmtSF', 'Heating',\n",
       "       'HeatingQC', 'CentralAir', 'Electrical', '1stFlrSF', '2ndFlrSF',\n",
       "       'LowQualFinSF', 'GrLivArea', 'BsmtFullBath', 'BsmtHalfBath', 'FullBath',\n",
       "       'HalfBath', 'BedroomAbvGr', 'KitchenAbvGr', 'KitchenQual',\n",
       "       'TotRmsAbvGrd', 'Functional', 'Fireplaces', 'FireplaceQu', 'GarageType',\n",
       "       'GarageYrBlt', 'GarageFinish', 'GarageCars', 'GarageArea', 'GarageQual',\n",
       "       'GarageCond', 'PavedDrive', 'WoodDeckSF', 'OpenPorchSF',\n",
       "       'EnclosedPorch', '3SsnPorch', 'ScreenPorch', 'PoolArea', 'PoolQC',\n",
       "       'Fence', 'MiscFeature', 'MiscVal', 'MoSold', 'YrSold', 'SaleType',\n",
       "       'SaleCondition', 'SalePrice'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 10;\n",
       "                var nbb_unformatted_code = \"train.columns\";\n",
       "                var nbb_formatted_code = \"train.columns\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.426247Z",
     "start_time": "2020-09-19T17:12:13.389928Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Building tailored column\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/work/.cache/poetry/xplainet-KS9Ipbc1-py3.7/lib/python3.7/site-packages/ipykernel_launcher.py:14: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \n",
      "/work/.cache/poetry/xplainet-KS9Ipbc1-py3.7/lib/python3.7/site-packages/pandas/core/indexing.py:670: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  iloc._setitem_with_indexer(indexer, value)\n",
      "/work/.cache/poetry/xplainet-KS9Ipbc1-py3.7/lib/python3.7/site-packages/ipykernel_launcher.py:15: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  from ipykernel import kernelapp as app\n"
     ]
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 11;\n",
       "                var nbb_unformatted_code = \"if \\\"Set\\\" not in train.columns:\\n    print(\\\"Building tailored column\\\")\\n    train_valid_index, test_index = next(\\n        ShuffleSplit(n_splits=1, test_size=0.1, random_state=SEED).split(\\n            range(train[target].shape[0])\\n        )\\n    )\\n    train_index, valid_index = next(\\n        ShuffleSplit(n_splits=1, test_size=0.1, random_state=SEED).split(\\n            train_valid_index\\n        )\\n    )\\n    train[\\\"Set\\\"] = \\\"train\\\"\\n    train[\\\"Set\\\"][valid_index] = \\\"valid\\\"\\n    train[\\\"Set\\\"][test_index] = \\\"test\\\"\\n    # train.to_csv((out.parent / \\\"train_bench.csv\\\").as_posix(), index=False)\";\n",
       "                var nbb_formatted_code = \"if \\\"Set\\\" not in train.columns:\\n    print(\\\"Building tailored column\\\")\\n    train_valid_index, test_index = next(\\n        ShuffleSplit(n_splits=1, test_size=0.1, random_state=SEED).split(\\n            range(train[target].shape[0])\\n        )\\n    )\\n    train_index, valid_index = next(\\n        ShuffleSplit(n_splits=1, test_size=0.1, random_state=SEED).split(\\n            train_valid_index\\n        )\\n    )\\n    train[\\\"Set\\\"] = \\\"train\\\"\\n    train[\\\"Set\\\"][valid_index] = \\\"valid\\\"\\n    train[\\\"Set\\\"][test_index] = \\\"test\\\"\\n    # train.to_csv((out.parent / \\\"train_bench.csv\\\").as_posix(), index=False)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if \"Set\" not in train.columns:\n",
    "    print(\"Building tailored column\")\n",
    "    train_valid_index, test_index = next(\n",
    "        ShuffleSplit(n_splits=1, test_size=0.1, random_state=SEED).split(\n",
    "            range(train[target].shape[0])\n",
    "        )\n",
    "    )\n",
    "    train_index, valid_index = next(\n",
    "        ShuffleSplit(n_splits=1, test_size=0.1, random_state=SEED).split(\n",
    "            train_valid_index\n",
    "        )\n",
    "    )\n",
    "    train[\"Set\"] = \"train\"\n",
    "    train[\"Set\"][valid_index] = \"valid\"\n",
    "    train[\"Set\"][test_index] = \"test\"\n",
    "    # train.to_csv((out.parent / \"train_bench.csv\").as_posix(), index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.457273Z",
     "start_time": "2020-09-19T17:12:13.427342Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 12;\n",
       "                var nbb_unformatted_code = \"train_indices = train[train.Set == \\\"train\\\"].index\\nvalid_indices = train[train.Set == \\\"valid\\\"].index\\ntest_indices = train[train.Set == \\\"test\\\"].index\";\n",
       "                var nbb_formatted_code = \"train_indices = train[train.Set == \\\"train\\\"].index\\nvalid_indices = train[train.Set == \\\"valid\\\"].index\\ntest_indices = train[train.Set == \\\"test\\\"].index\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_indices = train[train.Set == \"train\"].index\n",
    "valid_indices = train[train.Set == \"valid\"].index\n",
    "test_indices = train[train.Set == \"test\"].index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.604925Z",
     "start_time": "2020-09-19T17:12:13.458437Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'constant_cols': [],\n",
       " 'bool_cols': ['Alley', 'Street', 'CentralAir', 'Utilities'],\n",
       " 'num_cols': ['GarageYrBlt',\n",
       "  '1stFlrSF',\n",
       "  'LotFrontage',\n",
       "  'GrLivArea',\n",
       "  'ScreenPorch',\n",
       "  'TotalBsmtSF',\n",
       "  'BsmtFinSF1',\n",
       "  'Id',\n",
       "  'BsmtUnfSF',\n",
       "  'GarageArea',\n",
       "  'YearBuilt',\n",
       "  'BsmtFinSF2',\n",
       "  'WoodDeckSF',\n",
       "  'MasVnrArea',\n",
       "  'LotArea',\n",
       "  'EnclosedPorch',\n",
       "  '2ndFlrSF',\n",
       "  'YearRemodAdd',\n",
       "  'OpenPorchSF'],\n",
       " 'cat_cols': ['HeatingQC',\n",
       "  'Condition2',\n",
       "  'MiscFeature',\n",
       "  'RoofStyle',\n",
       "  'BsmtCond',\n",
       "  'Foundation',\n",
       "  'PavedDrive',\n",
       "  'LandSlope',\n",
       "  'MiscVal',\n",
       "  'LotConfig',\n",
       "  'GarageCond',\n",
       "  'ExterCond',\n",
       "  'SaleType',\n",
       "  'Fence',\n",
       "  'KitchenQual',\n",
       "  'SaleCondition',\n",
       "  'Fireplaces',\n",
       "  '3SsnPorch',\n",
       "  'GarageFinish',\n",
       "  'PoolQC',\n",
       "  'RoofMatl',\n",
       "  'BsmtHalfBath',\n",
       "  'GarageType',\n",
       "  'Heating',\n",
       "  'BsmtFullBath',\n",
       "  'Electrical',\n",
       "  'MSSubClass',\n",
       "  'Exterior1st',\n",
       "  'ExterQual',\n",
       "  'LandContour',\n",
       "  'TotRmsAbvGrd',\n",
       "  'LotShape',\n",
       "  'BsmtExposure',\n",
       "  'BedroomAbvGr',\n",
       "  'OverallQual',\n",
       "  'OverallCond',\n",
       "  'BsmtFinType2',\n",
       "  'Neighborhood',\n",
       "  'FullBath',\n",
       "  'Functional',\n",
       "  'FireplaceQu',\n",
       "  'MasVnrType',\n",
       "  'HalfBath',\n",
       "  'MoSold',\n",
       "  'YrSold',\n",
       "  'BsmtQual',\n",
       "  'BsmtFinType1',\n",
       "  'GarageCars',\n",
       "  'BldgType',\n",
       "  'MSZoning',\n",
       "  'HouseStyle',\n",
       "  'Exterior2nd',\n",
       "  'KitchenAbvGr',\n",
       "  'LowQualFinSF',\n",
       "  'Condition1',\n",
       "  'PoolArea',\n",
       "  'GarageQual'],\n",
       " 'num_encoder': [Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=1897.5279905941616,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=296.76122203505605,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=18.671047763804225,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=283.1521746236896,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-5.705547216467039,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-40.917873769333355,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-43.11290264214038,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-41.71908341967729,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-44.09528628925001,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-20.782345317596494,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=1868.954615177255,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-15.89740488044788,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-12.795001854935746,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-17.16005805073806,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=234.44552984495476,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-6.16716384502759,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-43.49782030326563,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=1947.9413545283148,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-6.805237977967029,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())])],\n",
       " 'bool_encoder': [SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder()],\n",
       " 'max_nb': 26,\n",
       " 'cat_encoder': [SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder()]}"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 13;\n",
       "                var nbb_unformatted_code = \"input_train, params = preproc_dataset(train.loc[train_indices], target, [\\\"Set\\\"])\\nparams\";\n",
       "                var nbb_formatted_code = \"input_train, params = preproc_dataset(train.loc[train_indices], target, [\\\"Set\\\"])\\nparams\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "input_train, params = preproc_dataset(train.loc[train_indices], target, [\"Set\"])\n",
    "params"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.633494Z",
     "start_time": "2020-09-19T17:12:13.605853Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1219"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 14;\n",
       "                var nbb_unformatted_code = \"len(train_indices)\";\n",
       "                var nbb_formatted_code = \"len(train_indices)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "len(train_indices)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.688683Z",
     "start_time": "2020-09-19T17:12:13.634346Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 15;\n",
       "                var nbb_unformatted_code = \"input_valid, _ = preproc_dataset(train.loc[valid_indices], target, [\\\"Set\\\"], params)\\ninput_test, _ = preproc_dataset(train.loc[test_indices], target, [\\\"Set\\\"], params)\";\n",
       "                var nbb_formatted_code = \"input_valid, _ = preproc_dataset(train.loc[valid_indices], target, [\\\"Set\\\"], params)\\ninput_test, _ = preproc_dataset(train.loc[test_indices], target, [\\\"Set\\\"], params)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "input_valid, _ = preproc_dataset(train.loc[valid_indices], target, [\"Set\"], params)\n",
    "input_test, _ = preproc_dataset(train.loc[test_indices], target, [\"Set\"], params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.712891Z",
     "start_time": "2020-09-19T17:12:13.689625Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 16;\n",
       "                var nbb_unformatted_code = \"target_encoder = StandardScaler()\";\n",
       "                var nbb_formatted_code = \"target_encoder = StandardScaler()\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "target_encoder = StandardScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.742312Z",
     "start_time": "2020-09-19T17:12:13.713791Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 17;\n",
       "                var nbb_unformatted_code = \"target_encoder.fit(train[target].values.reshape(-1, 1))\\ny_train = target_encoder.transform(train[target].values[train_indices].reshape(-1, 1))\\ny_valid = target_encoder.transform(train[target].values[valid_indices].reshape(-1, 1))\\ny_test = target_encoder.transform(train[target].values[test_indices].reshape(-1, 1))\";\n",
       "                var nbb_formatted_code = \"target_encoder.fit(train[target].values.reshape(-1, 1))\\ny_train = target_encoder.transform(train[target].values[train_indices].reshape(-1, 1))\\ny_valid = target_encoder.transform(train[target].values[valid_indices].reshape(-1, 1))\\ny_test = target_encoder.transform(train[target].values[test_indices].reshape(-1, 1))\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "target_encoder.fit(train[target].values.reshape(-1, 1))\n",
    "y_train = target_encoder.transform(train[target].values[train_indices].reshape(-1, 1))\n",
    "y_valid = target_encoder.transform(train[target].values[valid_indices].reshape(-1, 1))\n",
    "y_test = target_encoder.transform(train[target].values[test_indices].reshape(-1, 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:13.813317Z",
     "start_time": "2020-09-19T17:12:13.743550Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'constant_cols': [],\n",
       " 'bool_cols': ['Alley', 'Street', 'CentralAir', 'Utilities'],\n",
       " 'num_cols': ['GarageYrBlt',\n",
       "  '1stFlrSF',\n",
       "  'LotFrontage',\n",
       "  'GrLivArea',\n",
       "  'ScreenPorch',\n",
       "  'TotalBsmtSF',\n",
       "  'BsmtFinSF1',\n",
       "  'Id',\n",
       "  'BsmtUnfSF',\n",
       "  'GarageArea',\n",
       "  'YearBuilt',\n",
       "  'BsmtFinSF2',\n",
       "  'WoodDeckSF',\n",
       "  'MasVnrArea',\n",
       "  'LotArea',\n",
       "  'EnclosedPorch',\n",
       "  '2ndFlrSF',\n",
       "  'YearRemodAdd',\n",
       "  'OpenPorchSF'],\n",
       " 'cat_cols': ['HeatingQC',\n",
       "  'Condition2',\n",
       "  'MiscFeature',\n",
       "  'RoofStyle',\n",
       "  'BsmtCond',\n",
       "  'Foundation',\n",
       "  'PavedDrive',\n",
       "  'LandSlope',\n",
       "  'MiscVal',\n",
       "  'LotConfig',\n",
       "  'GarageCond',\n",
       "  'ExterCond',\n",
       "  'SaleType',\n",
       "  'Fence',\n",
       "  'KitchenQual',\n",
       "  'SaleCondition',\n",
       "  'Fireplaces',\n",
       "  '3SsnPorch',\n",
       "  'GarageFinish',\n",
       "  'PoolQC',\n",
       "  'RoofMatl',\n",
       "  'BsmtHalfBath',\n",
       "  'GarageType',\n",
       "  'Heating',\n",
       "  'BsmtFullBath',\n",
       "  'Electrical',\n",
       "  'MSSubClass',\n",
       "  'Exterior1st',\n",
       "  'ExterQual',\n",
       "  'LandContour',\n",
       "  'TotRmsAbvGrd',\n",
       "  'LotShape',\n",
       "  'BsmtExposure',\n",
       "  'BedroomAbvGr',\n",
       "  'OverallQual',\n",
       "  'OverallCond',\n",
       "  'BsmtFinType2',\n",
       "  'Neighborhood',\n",
       "  'FullBath',\n",
       "  'Functional',\n",
       "  'FireplaceQu',\n",
       "  'MasVnrType',\n",
       "  'HalfBath',\n",
       "  'MoSold',\n",
       "  'YrSold',\n",
       "  'BsmtQual',\n",
       "  'BsmtFinType1',\n",
       "  'GarageCars',\n",
       "  'BldgType',\n",
       "  'MSZoning',\n",
       "  'HouseStyle',\n",
       "  'Exterior2nd',\n",
       "  'KitchenAbvGr',\n",
       "  'LowQualFinSF',\n",
       "  'Condition1',\n",
       "  'PoolArea',\n",
       "  'GarageQual'],\n",
       " 'num_encoder': [Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=1897.5279905941616,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=296.76122203505605,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=18.671047763804225,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=283.1521746236896,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-5.705547216467039,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-40.917873769333355,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-43.11290264214038,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-41.71908341967729,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-44.09528628925001,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-20.782345317596494,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=1868.954615177255,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-15.89740488044788,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-12.795001854935746,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-17.16005805073806,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=234.44552984495476,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-6.16716384502759,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-43.49782030326563,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=1947.9413545283148,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())]),\n",
       "  Pipeline(steps=[('fillna',\n",
       "                   SimpleImputer(fill_value=-6.805237977967029,\n",
       "                                 strategy='constant')),\n",
       "                  ('scaler', StandardScaler())])],\n",
       " 'bool_encoder': [SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder()],\n",
       " 'max_nb': 26,\n",
       " 'cat_encoder': [SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder(),\n",
       "  SafeLabelEncoder()]}"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 18;\n",
       "                var nbb_unformatted_code = \"params\";\n",
       "                var nbb_formatted_code = \"params\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "params"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:14.037016Z",
     "start_time": "2020-09-19T17:12:13.814205Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/work/.cache/poetry/xplainet-KS9Ipbc1-py3.7/lib/python3.7/site-packages/tensorflow_addons/utils/resource_loader.py:103: UserWarning: You are currently using TensorFlow 2.3.0 and trying to load a custom op (custom_ops/activations/_activation_ops.so).\n",
      "TensorFlow Addons has compiled its custom ops against TensorFlow 2.2.0, and there are no compatibility guarantees between the two versions. \n",
      "This means that you might get segfaults when loading the custom op, or other kind of low-level errors.\n",
      " If you do, do not file an issue on Github. This is a known limitation.\n",
      "\n",
      "It might help you to fallback to pure Python ops with TF_ADDONS_PY_OPS . To do that, see https://github.com/tensorflow/addons#gpucpu-custom-ops \n",
      "\n",
      "You can also change the TensorFlow version installed on your system. You would need a TensorFlow version equal to or above 2.2.0 and strictly below 2.3.0.\n",
      " Note that nightly versions of TensorFlow, as well as non-pip TensorFlow like `conda install tensorflow` or compiled from source are not supported.\n",
      "\n",
      "The last solution is to find the TensorFlow Addons version that has custom ops compatible with the TensorFlow installed on your system. To do that, refer to the readme: https://github.com/tensorflow/addons\n",
      "  UserWarning,\n"
     ]
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 19;\n",
       "                var nbb_unformatted_code = \"model = build_model(\\n    params,\\n    lconv_dim=[4],\\n    lconv_num_dim=[8],\\n    emb_size=16,\\n    # For this problem, we need \\\"tanh\\\" as first layer, or else to standard scale the data beforehand\\n    activation_num_first_layer=\\\"tanh\\\",\\n    output_activation=None,\\n    output_dim=1,  # np.unique(y_train).shape[0],\\n)\";\n",
       "                var nbb_formatted_code = \"model = build_model(\\n    params,\\n    lconv_dim=[4],\\n    lconv_num_dim=[8],\\n    emb_size=16,\\n    # For this problem, we need \\\"tanh\\\" as first layer, or else to standard scale the data beforehand\\n    activation_num_first_layer=\\\"tanh\\\",\\n    output_activation=None,\\n    output_dim=1,  # np.unique(y_train).shape[0],\\n)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = build_model(\n",
    "    params,\n",
    "    lconv_dim=[4],\n",
    "    lconv_num_dim=[8],\n",
    "    emb_size=16,\n",
    "    # For this problem, we need \"tanh\" as first layer, or else to standard scale the data beforehand\n",
    "    activation_num_first_layer=\"tanh\",\n",
    "    output_activation=None,\n",
    "    output_dim=1,  # np.unique(y_train).shape[0],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:14.064317Z",
     "start_time": "2020-09-19T17:12:14.038066Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"explainable_model\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_num (InputLayer)          [(None, 19)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "input_cat (InputLayer)          [(None, 57)]         0                                            \n",
      "__________________________________________________________________________________________________\n",
      "reshape_num_input (Reshape)     (None, 19, 1)        0           input_num[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "large_emb (Embedding)           (None, 57, 16)       432         input_cat[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "block_num_0_conv (LocallyConnec (None, 19, 8)        152         reshape_num_input[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "block_cat_0_conv (LocallyConnec (None, 57, 4)        3648        large_emb[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "block_num_0_activation (Activat (None, 19, 8)        0           block_num_0_conv[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "block_cat_0_activation (Activat (None, 57, 4)        0           block_cat_0_conv[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "input_bool (InputLayer)         [(None, 4)]          0                                            \n",
      "__________________________________________________________________________________________________\n",
      "reshape_num_output (Reshape)    (None, 152)          0           block_num_0_activation[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "reshape_cat_output (Reshape)    (None, 228)          0           block_cat_0_activation[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "concatenate (Concatenate)       (None, 384)          0           input_bool[0][0]                 \n",
      "                                                                 reshape_num_output[0][0]         \n",
      "                                                                 reshape_cat_output[0][0]         \n",
      "__________________________________________________________________________________________________\n",
      "output (Dense)                  (None, 1)            385         concatenate[0][0]                \n",
      "==================================================================================================\n",
      "Total params: 4,617\n",
      "Trainable params: 4,617\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 20;\n",
       "                var nbb_unformatted_code = \"model.summary()\";\n",
       "                var nbb_formatted_code = \"model.summary()\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:14.093382Z",
     "start_time": "2020-09-19T17:12:14.065557Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 21;\n",
       "                var nbb_unformatted_code = \"#!pip install pydot graphviz\";\n",
       "                var nbb_formatted_code = \"#!pip install pydot graphviz\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#!pip install pydot graphviz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:14.304007Z",
     "start_time": "2020-09-19T17:12:14.094898Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 22;\n",
       "                var nbb_unformatted_code = \"plot_model(\\n    model,\\n    # to_file=\\\"model.png\\\",\\n    show_shapes=True,\\n    show_layer_names=True,\\n    rankdir=\\\"TB\\\",\\n    expand_nested=False,\\n    dpi=96,\\n)\";\n",
       "                var nbb_formatted_code = \"plot_model(\\n    model,\\n    # to_file=\\\"model.png\\\",\\n    show_shapes=True,\\n    show_layer_names=True,\\n    rankdir=\\\"TB\\\",\\n    expand_nested=False,\\n    dpi=96,\\n)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_model(\n",
    "    model,\n",
    "    # to_file=\"model.png\",\n",
    "    show_shapes=True,\n",
    "    show_layer_names=True,\n",
    "    rankdir=\"TB\",\n",
    "    expand_nested=False,\n",
    "    dpi=96,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:14.341150Z",
     "start_time": "2020-09-19T17:12:14.307707Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1219, 1)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 23;\n",
       "                var nbb_unformatted_code = \"y_train.shape\";\n",
       "                var nbb_formatted_code = \"y_train.shape\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:14.371760Z",
     "start_time": "2020-09-19T17:12:14.343340Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 24;\n",
       "                var nbb_unformatted_code = \"counts = np.unique(y_train, return_counts=True)[1]\\ncounts = counts.sum() / counts\";\n",
       "                var nbb_formatted_code = \"counts = np.unique(y_train, return_counts=True)[1]\\ncounts = counts.sum() / counts\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "counts = np.unique(y_train, return_counts=True)[1]\n",
    "counts = counts.sum() / counts"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T11:09:21.280303Z",
     "start_time": "2020-09-19T11:09:21.244198Z"
    }
   },
   "source": [
    "class_weight = {}\n",
    "for i, counts in enumerate(counts):\n",
    "    class_weight[i] = counts\n",
    "class_weight"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-14T10:13:28.824624Z",
     "start_time": "2020-07-14T10:13:28.800384Z"
    }
   },
   "source": [
    "class_weight = {\n",
    "    0: 1,\n",
    "    1: 1,\n",
    "}\n",
    "class_weight"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:34.400331Z",
     "start_time": "2020-09-19T17:12:14.372769Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/2000\n",
      "2/2 - 0s - loss: 0.9952 - val_loss: 1.0720\n",
      "Epoch 2/2000\n",
      "2/2 - 0s - loss: 0.9901 - val_loss: 1.0669\n",
      "Epoch 3/2000\n",
      "2/2 - 0s - loss: 0.9854 - val_loss: 1.0705\n",
      "Epoch 4/2000\n",
      "2/2 - 0s - loss: 0.9890 - val_loss: 1.0689\n",
      "Epoch 5/2000\n",
      "2/2 - 0s - loss: 0.9876 - val_loss: 1.0669\n",
      "Epoch 6/2000\n",
      "2/2 - 0s - loss: 0.9857 - val_loss: 1.0675\n",
      "Epoch 7/2000\n",
      "2/2 - 0s - loss: 0.9862 - val_loss: 1.0647\n",
      "Epoch 8/2000\n",
      "2/2 - 0s - loss: 0.9837 - val_loss: 1.0616\n",
      "Epoch 9/2000\n",
      "2/2 - 0s - loss: 0.9809 - val_loss: 1.0629\n",
      "Epoch 10/2000\n",
      "2/2 - 0s - loss: 0.9820 - val_loss: 1.0593\n",
      "Epoch 11/2000\n",
      "2/2 - 0s - loss: 0.9787 - val_loss: 1.0554\n",
      "Epoch 12/2000\n",
      "2/2 - 0s - loss: 0.9754 - val_loss: 1.0571\n",
      "Epoch 13/2000\n",
      "2/2 - 0s - loss: 0.9768 - val_loss: 1.0529\n",
      "Epoch 14/2000\n",
      "2/2 - 0s - loss: 0.9730 - val_loss: 1.0485\n",
      "Epoch 15/2000\n",
      "2/2 - 0s - loss: 0.9690 - val_loss: 1.0505\n",
      "Epoch 16/2000\n",
      "2/2 - 0s - loss: 0.9708 - val_loss: 1.0456\n",
      "Epoch 17/2000\n",
      "2/2 - 0s - loss: 0.9665 - val_loss: 1.0407\n",
      "Epoch 18/2000\n",
      "2/2 - 0s - loss: 0.9620 - val_loss: 1.0430\n",
      "Epoch 19/2000\n",
      "2/2 - 0s - loss: 0.9640 - val_loss: 1.0377\n",
      "Epoch 20/2000\n",
      "2/2 - 0s - loss: 0.9592 - val_loss: 1.0323\n",
      "Epoch 21/2000\n",
      "2/2 - 0s - loss: 0.9543 - val_loss: 1.0348\n",
      "Epoch 22/2000\n",
      "2/2 - 0s - loss: 0.9567 - val_loss: 1.0291\n",
      "Epoch 23/2000\n",
      "2/2 - 0s - loss: 0.9515 - val_loss: 1.0233\n",
      "Epoch 24/2000\n",
      "2/2 - 0s - loss: 0.9462 - val_loss: 1.0261\n",
      "Epoch 25/2000\n",
      "2/2 - 0s - loss: 0.9487 - val_loss: 1.0200\n",
      "Epoch 26/2000\n",
      "2/2 - 0s - loss: 0.9432 - val_loss: 1.0137\n",
      "Epoch 27/2000\n",
      "2/2 - 0s - loss: 0.9377 - val_loss: 1.0167\n",
      "Epoch 28/2000\n",
      "2/2 - 0s - loss: 0.9403 - val_loss: 1.0103\n",
      "Epoch 29/2000\n",
      "2/2 - 0s - loss: 0.9345 - val_loss: 1.0038\n",
      "Epoch 30/2000\n",
      "2/2 - 0s - loss: 0.9287 - val_loss: 1.0070\n",
      "Epoch 31/2000\n",
      "2/2 - 0s - loss: 0.9314 - val_loss: 1.0003\n",
      "Epoch 32/2000\n",
      "2/2 - 0s - loss: 0.9254 - val_loss: 0.9935\n",
      "Epoch 33/2000\n",
      "2/2 - 0s - loss: 0.9193 - val_loss: 0.9968\n",
      "Epoch 34/2000\n",
      "2/2 - 0s - loss: 0.9223 - val_loss: 0.9899\n",
      "Epoch 35/2000\n",
      "2/2 - 0s - loss: 0.9159 - val_loss: 0.9829\n",
      "Epoch 36/2000\n",
      "2/2 - 0s - loss: 0.9097 - val_loss: 0.9863\n",
      "Epoch 37/2000\n",
      "2/2 - 0s - loss: 0.9128 - val_loss: 0.9792\n",
      "Epoch 38/2000\n",
      "2/2 - 0s - loss: 0.9063 - val_loss: 0.9721\n",
      "Epoch 39/2000\n",
      "2/2 - 0s - loss: 0.8998 - val_loss: 0.9756\n",
      "Epoch 40/2000\n",
      "2/2 - 0s - loss: 0.9028 - val_loss: 0.9682\n",
      "Epoch 41/2000\n",
      "2/2 - 0s - loss: 0.8963 - val_loss: 0.9608\n",
      "Epoch 42/2000\n",
      "2/2 - 0s - loss: 0.8895 - val_loss: 0.9644\n",
      "Epoch 43/2000\n",
      "2/2 - 0s - loss: 0.8927 - val_loss: 0.9568\n",
      "Epoch 44/2000\n",
      "2/2 - 0s - loss: 0.8859 - val_loss: 0.9492\n",
      "Epoch 45/2000\n",
      "2/2 - 0s - loss: 0.8790 - val_loss: 0.9529\n",
      "Epoch 46/2000\n",
      "2/2 - 0s - loss: 0.8823 - val_loss: 0.9452\n",
      "Epoch 47/2000\n",
      "2/2 - 0s - loss: 0.8752 - val_loss: 0.9375\n",
      "Epoch 48/2000\n",
      "2/2 - 0s - loss: 0.8681 - val_loss: 0.9412\n",
      "Epoch 49/2000\n",
      "2/2 - 0s - loss: 0.8717 - val_loss: 0.9333\n",
      "Epoch 50/2000\n",
      "2/2 - 0s - loss: 0.8644 - val_loss: 0.9254\n",
      "Epoch 51/2000\n",
      "2/2 - 0s - loss: 0.8570 - val_loss: 0.9293\n",
      "Epoch 52/2000\n",
      "2/2 - 0s - loss: 0.8608 - val_loss: 0.9212\n",
      "Epoch 53/2000\n",
      "2/2 - 0s - loss: 0.8533 - val_loss: 0.9132\n",
      "Epoch 54/2000\n",
      "2/2 - 0s - loss: 0.8460 - val_loss: 0.9172\n",
      "Epoch 55/2000\n",
      "2/2 - 0s - loss: 0.8496 - val_loss: 0.9090\n",
      "Epoch 56/2000\n",
      "2/2 - 0s - loss: 0.8423 - val_loss: 0.9009\n",
      "Epoch 57/2000\n",
      "2/2 - 0s - loss: 0.8347 - val_loss: 0.9049\n",
      "Epoch 58/2000\n",
      "2/2 - 0s - loss: 0.8384 - val_loss: 0.8966\n",
      "Epoch 59/2000\n",
      "2/2 - 0s - loss: 0.8309 - val_loss: 0.8884\n",
      "Epoch 60/2000\n",
      "2/2 - 0s - loss: 0.8233 - val_loss: 0.8925\n",
      "Epoch 61/2000\n",
      "2/2 - 0s - loss: 0.8270 - val_loss: 0.8841\n",
      "Epoch 62/2000\n",
      "2/2 - 0s - loss: 0.8194 - val_loss: 0.8758\n",
      "Epoch 63/2000\n",
      "2/2 - 0s - loss: 0.8118 - val_loss: 0.8799\n",
      "Epoch 64/2000\n",
      "2/2 - 0s - loss: 0.8154 - val_loss: 0.8714\n",
      "Epoch 65/2000\n",
      "2/2 - 0s - loss: 0.8076 - val_loss: 0.8628\n",
      "Epoch 66/2000\n",
      "2/2 - 0s - loss: 0.7996 - val_loss: 0.8669\n",
      "Epoch 67/2000\n",
      "2/2 - 0s - loss: 0.8035 - val_loss: 0.8580\n",
      "Epoch 68/2000\n",
      "2/2 - 0s - loss: 0.7955 - val_loss: 0.8492\n",
      "Epoch 69/2000\n",
      "2/2 - 0s - loss: 0.7873 - val_loss: 0.8536\n",
      "Epoch 70/2000\n",
      "2/2 - 0s - loss: 0.7914 - val_loss: 0.8449\n",
      "Epoch 71/2000\n",
      "2/2 - 0s - loss: 0.7833 - val_loss: 0.8361\n",
      "Epoch 72/2000\n",
      "2/2 - 0s - loss: 0.7752 - val_loss: 0.8404\n",
      "Epoch 73/2000\n",
      "2/2 - 0s - loss: 0.7792 - val_loss: 0.8315\n",
      "Epoch 74/2000\n",
      "2/2 - 0s - loss: 0.7709 - val_loss: 0.8226\n",
      "Epoch 75/2000\n",
      "2/2 - 0s - loss: 0.7626 - val_loss: 0.8269\n",
      "Epoch 76/2000\n",
      "2/2 - 0s - loss: 0.7668 - val_loss: 0.8179\n",
      "Epoch 77/2000\n",
      "2/2 - 0s - loss: 0.7583 - val_loss: 0.8089\n",
      "Epoch 78/2000\n",
      "2/2 - 0s - loss: 0.7500 - val_loss: 0.8134\n",
      "Epoch 79/2000\n",
      "2/2 - 0s - loss: 0.7541 - val_loss: 0.8044\n",
      "Epoch 80/2000\n",
      "2/2 - 0s - loss: 0.7456 - val_loss: 0.7953\n",
      "Epoch 81/2000\n",
      "2/2 - 0s - loss: 0.7371 - val_loss: 0.7997\n",
      "Epoch 82/2000\n",
      "2/2 - 0s - loss: 0.7412 - val_loss: 0.7904\n",
      "Epoch 83/2000\n",
      "2/2 - 0s - loss: 0.7325 - val_loss: 0.7812\n",
      "Epoch 84/2000\n",
      "2/2 - 0s - loss: 0.7240 - val_loss: 0.7857\n",
      "Epoch 85/2000\n",
      "2/2 - 0s - loss: 0.7281 - val_loss: 0.7765\n",
      "Epoch 86/2000\n",
      "2/2 - 0s - loss: 0.7195 - val_loss: 0.7671\n",
      "Epoch 87/2000\n",
      "2/2 - 0s - loss: 0.7107 - val_loss: 0.7717\n",
      "Epoch 88/2000\n",
      "2/2 - 0s - loss: 0.7149 - val_loss: 0.7623\n",
      "Epoch 89/2000\n",
      "2/2 - 0s - loss: 0.7061 - val_loss: 0.7528\n",
      "Epoch 90/2000\n",
      "2/2 - 0s - loss: 0.6971 - val_loss: 0.7575\n",
      "Epoch 91/2000\n",
      "2/2 - 0s - loss: 0.7016 - val_loss: 0.7480\n",
      "Epoch 92/2000\n",
      "2/2 - 0s - loss: 0.6926 - val_loss: 0.7386\n",
      "Epoch 93/2000\n",
      "2/2 - 0s - loss: 0.6836 - val_loss: 0.7433\n",
      "Epoch 94/2000\n",
      "2/2 - 0s - loss: 0.6880 - val_loss: 0.7338\n",
      "Epoch 95/2000\n",
      "2/2 - 0s - loss: 0.6789 - val_loss: 0.7243\n",
      "Epoch 96/2000\n",
      "2/2 - 0s - loss: 0.6699 - val_loss: 0.7290\n",
      "Epoch 97/2000\n",
      "2/2 - 0s - loss: 0.6743 - val_loss: 0.7193\n",
      "Epoch 98/2000\n",
      "2/2 - 0s - loss: 0.6652 - val_loss: 0.7096\n",
      "Epoch 99/2000\n",
      "2/2 - 0s - loss: 0.6560 - val_loss: 0.7144\n",
      "Epoch 100/2000\n",
      "2/2 - 0s - loss: 0.6604 - val_loss: 0.7047\n",
      "Epoch 101/2000\n",
      "2/2 - 0s - loss: 0.6511 - val_loss: 0.6948\n",
      "Epoch 102/2000\n",
      "2/2 - 0s - loss: 0.6417 - val_loss: 0.6996\n",
      "Epoch 103/2000\n",
      "2/2 - 0s - loss: 0.6466 - val_loss: 0.6896\n",
      "Epoch 104/2000\n",
      "2/2 - 0s - loss: 0.6371 - val_loss: 0.6798\n",
      "Epoch 105/2000\n",
      "2/2 - 0s - loss: 0.6275 - val_loss: 0.6848\n",
      "Epoch 106/2000\n",
      "2/2 - 0s - loss: 0.6322 - val_loss: 0.6750\n",
      "Epoch 107/2000\n",
      "2/2 - 0s - loss: 0.6228 - val_loss: 0.6651\n",
      "Epoch 108/2000\n",
      "2/2 - 0s - loss: 0.6131 - val_loss: 0.6699\n",
      "Epoch 109/2000\n",
      "2/2 - 0s - loss: 0.6181 - val_loss: 0.6600\n",
      "Epoch 110/2000\n",
      "2/2 - 0s - loss: 0.6081 - val_loss: 0.6502\n",
      "Epoch 111/2000\n",
      "2/2 - 0s - loss: 0.5985 - val_loss: 0.6551\n",
      "Epoch 112/2000\n",
      "2/2 - 0s - loss: 0.6034 - val_loss: 0.6451\n",
      "Epoch 113/2000\n",
      "2/2 - 0s - loss: 0.5936 - val_loss: 0.6351\n",
      "Epoch 114/2000\n",
      "2/2 - 0s - loss: 0.5839 - val_loss: 0.6401\n",
      "Epoch 115/2000\n",
      "2/2 - 0s - loss: 0.5887 - val_loss: 0.6301\n",
      "Epoch 116/2000\n",
      "2/2 - 0s - loss: 0.5790 - val_loss: 0.6201\n",
      "Epoch 117/2000\n",
      "2/2 - 0s - loss: 0.5692 - val_loss: 0.6251\n",
      "Epoch 118/2000\n",
      "2/2 - 0s - loss: 0.5740 - val_loss: 0.6152\n",
      "Epoch 119/2000\n",
      "2/2 - 0s - loss: 0.5643 - val_loss: 0.6053\n",
      "Epoch 120/2000\n",
      "2/2 - 0s - loss: 0.5543 - val_loss: 0.6102\n",
      "Epoch 121/2000\n",
      "2/2 - 0s - loss: 0.5594 - val_loss: 0.6001\n",
      "Epoch 122/2000\n",
      "2/2 - 0s - loss: 0.5494 - val_loss: 0.5901\n",
      "Epoch 123/2000\n",
      "2/2 - 0s - loss: 0.5394 - val_loss: 0.5952\n",
      "Epoch 124/2000\n",
      "2/2 - 0s - loss: 0.5444 - val_loss: 0.5852\n",
      "Epoch 125/2000\n",
      "2/2 - 0s - loss: 0.5345 - val_loss: 0.5753\n",
      "Epoch 126/2000\n",
      "2/2 - 0s - loss: 0.5245 - val_loss: 0.5802\n",
      "Epoch 127/2000\n",
      "2/2 - 0s - loss: 0.5294 - val_loss: 0.5701\n",
      "Epoch 128/2000\n",
      "2/2 - 0s - loss: 0.5193 - val_loss: 0.5600\n",
      "Epoch 129/2000\n",
      "2/2 - 0s - loss: 0.5092 - val_loss: 0.5650\n",
      "Epoch 130/2000\n",
      "2/2 - 0s - loss: 0.5141 - val_loss: 0.5549\n",
      "Epoch 131/2000\n",
      "2/2 - 0s - loss: 0.5041 - val_loss: 0.5447\n",
      "Epoch 132/2000\n",
      "2/2 - 0s - loss: 0.4937 - val_loss: 0.5498\n",
      "Epoch 133/2000\n",
      "2/2 - 0s - loss: 0.4988 - val_loss: 0.5397\n",
      "Epoch 134/2000\n",
      "2/2 - 0s - loss: 0.4886 - val_loss: 0.5296\n",
      "Epoch 135/2000\n",
      "2/2 - 0s - loss: 0.4785 - val_loss: 0.5346\n",
      "Epoch 136/2000\n",
      "2/2 - 0s - loss: 0.4837 - val_loss: 0.5247\n",
      "Epoch 137/2000\n",
      "2/2 - 0s - loss: 0.4734 - val_loss: 0.5149\n",
      "Epoch 138/2000\n",
      "2/2 - 0s - loss: 0.4634 - val_loss: 0.5199\n",
      "Epoch 139/2000\n",
      "2/2 - 0s - loss: 0.4683 - val_loss: 0.5102\n",
      "Epoch 140/2000\n",
      "2/2 - 0s - loss: 0.4583 - val_loss: 0.5005\n",
      "Epoch 141/2000\n",
      "2/2 - 0s - loss: 0.4482 - val_loss: 0.5052\n",
      "Epoch 142/2000\n",
      "2/2 - 0s - loss: 0.4530 - val_loss: 0.4953\n",
      "Epoch 143/2000\n",
      "2/2 - 0s - loss: 0.4429 - val_loss: 0.4854\n",
      "Epoch 144/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2/2 - 0s - loss: 0.4324 - val_loss: 0.4903\n",
      "Epoch 145/2000\n",
      "2/2 - 0s - loss: 0.4373 - val_loss: 0.4803\n",
      "Epoch 146/2000\n",
      "2/2 - 0s - loss: 0.4271 - val_loss: 0.4703\n",
      "Epoch 147/2000\n",
      "2/2 - 0s - loss: 0.4168 - val_loss: 0.4753\n",
      "Epoch 148/2000\n",
      "2/2 - 0s - loss: 0.4218 - val_loss: 0.4656\n",
      "Epoch 149/2000\n",
      "2/2 - 0s - loss: 0.4117 - val_loss: 0.4561\n",
      "Epoch 150/2000\n",
      "2/2 - 0s - loss: 0.4017 - val_loss: 0.4609\n",
      "Epoch 151/2000\n",
      "2/2 - 0s - loss: 0.4066 - val_loss: 0.4517\n",
      "Epoch 152/2000\n",
      "2/2 - 0s - loss: 0.3968 - val_loss: 0.4425\n",
      "Epoch 153/2000\n",
      "2/2 - 0s - loss: 0.3871 - val_loss: 0.4471\n",
      "Epoch 154/2000\n",
      "2/2 - 0s - loss: 0.3919 - val_loss: 0.4381\n",
      "Epoch 155/2000\n",
      "2/2 - 0s - loss: 0.3822 - val_loss: 0.4294\n",
      "Epoch 156/2000\n",
      "2/2 - 0s - loss: 0.3731 - val_loss: 0.4338\n",
      "Epoch 157/2000\n",
      "2/2 - 0s - loss: 0.3778 - val_loss: 0.4257\n",
      "Epoch 158/2000\n",
      "2/2 - 0s - loss: 0.3686 - val_loss: 0.4179\n",
      "Epoch 159/2000\n",
      "2/2 - 0s - loss: 0.3599 - val_loss: 0.4218\n",
      "Epoch 160/2000\n",
      "2/2 - 0s - loss: 0.3644 - val_loss: 0.4143\n",
      "Epoch 161/2000\n",
      "2/2 - 0s - loss: 0.3556 - val_loss: 0.4069\n",
      "Epoch 162/2000\n",
      "2/2 - 0s - loss: 0.3472 - val_loss: 0.4105\n",
      "Epoch 163/2000\n",
      "2/2 - 0s - loss: 0.3514 - val_loss: 0.4032\n",
      "Epoch 164/2000\n",
      "2/2 - 0s - loss: 0.3428 - val_loss: 0.3960\n",
      "Epoch 165/2000\n",
      "2/2 - 0s - loss: 0.3345 - val_loss: 0.3996\n",
      "Epoch 166/2000\n",
      "2/2 - 0s - loss: 0.3388 - val_loss: 0.3925\n",
      "Epoch 167/2000\n",
      "2/2 - 0s - loss: 0.3305 - val_loss: 0.3858\n",
      "Epoch 168/2000\n",
      "2/2 - 0s - loss: 0.3224 - val_loss: 0.3892\n",
      "Epoch 169/2000\n",
      "2/2 - 0s - loss: 0.3267 - val_loss: 0.3827\n",
      "Epoch 170/2000\n",
      "2/2 - 0s - loss: 0.3188 - val_loss: 0.3766\n",
      "Epoch 171/2000\n",
      "2/2 - 0s - loss: 0.3113 - val_loss: 0.3797\n",
      "Epoch 172/2000\n",
      "2/2 - 0s - loss: 0.3150 - val_loss: 0.3737\n",
      "Epoch 173/2000\n",
      "2/2 - 0s - loss: 0.3080 - val_loss: 0.3682\n",
      "Epoch 174/2000\n",
      "2/2 - 0s - loss: 0.3007 - val_loss: 0.3711\n",
      "Epoch 175/2000\n",
      "2/2 - 0s - loss: 0.3043 - val_loss: 0.3660\n",
      "Epoch 176/2000\n",
      "2/2 - 0s - loss: 0.2976 - val_loss: 0.3610\n",
      "Epoch 177/2000\n",
      "2/2 - 0s - loss: 0.2910 - val_loss: 0.3635\n",
      "Epoch 178/2000\n",
      "2/2 - 0s - loss: 0.2941 - val_loss: 0.3586\n",
      "Epoch 179/2000\n",
      "2/2 - 0s - loss: 0.2879 - val_loss: 0.3539\n",
      "Epoch 180/2000\n",
      "2/2 - 0s - loss: 0.2815 - val_loss: 0.3563\n",
      "Epoch 181/2000\n",
      "2/2 - 0s - loss: 0.2846 - val_loss: 0.3519\n",
      "Epoch 182/2000\n",
      "2/2 - 0s - loss: 0.2785 - val_loss: 0.3477\n",
      "Epoch 183/2000\n",
      "2/2 - 0s - loss: 0.2725 - val_loss: 0.3497\n",
      "Epoch 184/2000\n",
      "2/2 - 0s - loss: 0.2756 - val_loss: 0.3457\n",
      "Epoch 185/2000\n",
      "2/2 - 0s - loss: 0.2697 - val_loss: 0.3419\n",
      "Epoch 186/2000\n",
      "2/2 - 0s - loss: 0.2642 - val_loss: 0.3438\n",
      "Epoch 187/2000\n",
      "2/2 - 0s - loss: 0.2671 - val_loss: 0.3402\n",
      "Epoch 188/2000\n",
      "2/2 - 0s - loss: 0.2614 - val_loss: 0.3368\n",
      "Epoch 189/2000\n",
      "2/2 - 0s - loss: 0.2565 - val_loss: 0.3385\n",
      "Epoch 190/2000\n",
      "2/2 - 0s - loss: 0.2593 - val_loss: 0.3353\n",
      "Epoch 191/2000\n",
      "2/2 - 0s - loss: 0.2538 - val_loss: 0.3325\n",
      "Epoch 192/2000\n",
      "2/2 - 0s - loss: 0.2493 - val_loss: 0.3339\n",
      "Epoch 193/2000\n",
      "2/2 - 0s - loss: 0.2516 - val_loss: 0.3312\n",
      "Epoch 194/2000\n",
      "2/2 - 0s - loss: 0.2470 - val_loss: 0.3286\n",
      "Epoch 195/2000\n",
      "2/2 - 0s - loss: 0.2425 - val_loss: 0.3299\n",
      "Epoch 196/2000\n",
      "2/2 - 0s - loss: 0.2448 - val_loss: 0.3276\n",
      "Epoch 197/2000\n",
      "2/2 - 0s - loss: 0.2405 - val_loss: 0.3255\n",
      "Epoch 198/2000\n",
      "2/2 - 0s - loss: 0.2364 - val_loss: 0.3265\n",
      "Epoch 199/2000\n",
      "2/2 - 0s - loss: 0.2383 - val_loss: 0.3246\n",
      "Epoch 200/2000\n",
      "2/2 - 0s - loss: 0.2344 - val_loss: 0.3227\n",
      "Epoch 201/2000\n",
      "2/2 - 0s - loss: 0.2304 - val_loss: 0.3236\n",
      "Epoch 202/2000\n",
      "2/2 - 0s - loss: 0.2324 - val_loss: 0.3219\n",
      "Epoch 203/2000\n",
      "2/2 - 0s - loss: 0.2288 - val_loss: 0.3204\n",
      "Epoch 204/2000\n",
      "2/2 - 0s - loss: 0.2252 - val_loss: 0.3212\n",
      "Epoch 205/2000\n",
      "2/2 - 0s - loss: 0.2270 - val_loss: 0.3200\n",
      "Epoch 206/2000\n",
      "2/2 - 0s - loss: 0.2239 - val_loss: 0.3189\n",
      "Epoch 207/2000\n",
      "2/2 - 0s - loss: 0.2205 - val_loss: 0.3194\n",
      "Epoch 208/2000\n",
      "2/2 - 0s - loss: 0.2223 - val_loss: 0.3184\n",
      "Epoch 209/2000\n",
      "2/2 - 0s - loss: 0.2191 - val_loss: 0.3175\n",
      "Epoch 210/2000\n",
      "2/2 - 0s - loss: 0.2161 - val_loss: 0.3179\n",
      "Epoch 211/2000\n",
      "2/2 - 0s - loss: 0.2177 - val_loss: 0.3171\n",
      "Epoch 212/2000\n",
      "2/2 - 0s - loss: 0.2149 - val_loss: 0.3164\n",
      "Epoch 213/2000\n",
      "2/2 - 0s - loss: 0.2121 - val_loss: 0.3168\n",
      "Epoch 214/2000\n",
      "2/2 - 0s - loss: 0.2137 - val_loss: 0.3162\n",
      "Epoch 215/2000\n",
      "2/2 - 0s - loss: 0.2109 - val_loss: 0.3157\n",
      "Epoch 216/2000\n",
      "2/2 - 0s - loss: 0.2086 - val_loss: 0.3160\n",
      "Epoch 217/2000\n",
      "2/2 - 0s - loss: 0.2098 - val_loss: 0.3156\n",
      "Epoch 218/2000\n",
      "2/2 - 0s - loss: 0.2075 - val_loss: 0.3152\n",
      "Epoch 219/2000\n",
      "2/2 - 0s - loss: 0.2052 - val_loss: 0.3154\n",
      "Epoch 220/2000\n",
      "2/2 - 0s - loss: 0.2064 - val_loss: 0.3149\n",
      "Epoch 221/2000\n",
      "2/2 - 0s - loss: 0.2041 - val_loss: 0.3146\n",
      "Epoch 222/2000\n",
      "2/2 - 0s - loss: 0.2020 - val_loss: 0.3148\n",
      "Epoch 223/2000\n",
      "2/2 - 0s - loss: 0.2030 - val_loss: 0.3145\n",
      "Epoch 224/2000\n",
      "2/2 - 0s - loss: 0.2008 - val_loss: 0.3142\n",
      "Epoch 225/2000\n",
      "2/2 - 0s - loss: 0.1988 - val_loss: 0.3143\n",
      "Epoch 226/2000\n",
      "2/2 - 0s - loss: 0.1997 - val_loss: 0.3141\n",
      "Epoch 227/2000\n",
      "2/2 - 0s - loss: 0.1978 - val_loss: 0.3140\n",
      "Epoch 228/2000\n",
      "2/2 - 0s - loss: 0.1957 - val_loss: 0.3140\n",
      "Epoch 229/2000\n",
      "2/2 - 0s - loss: 0.1967 - val_loss: 0.3139\n",
      "Epoch 230/2000\n",
      "2/2 - 0s - loss: 0.1948 - val_loss: 0.3139\n",
      "Epoch 231/2000\n",
      "2/2 - 0s - loss: 0.1928 - val_loss: 0.3139\n",
      "Epoch 232/2000\n",
      "2/2 - 0s - loss: 0.1939 - val_loss: 0.3138\n",
      "Epoch 233/2000\n",
      "2/2 - 0s - loss: 0.1918 - val_loss: 0.3138\n",
      "Epoch 234/2000\n",
      "2/2 - 0s - loss: 0.1902 - val_loss: 0.3138\n",
      "Epoch 235/2000\n",
      "2/2 - 0s - loss: 0.1910 - val_loss: 0.3137\n",
      "Epoch 236/2000\n",
      "2/2 - 0s - loss: 0.1892 - val_loss: 0.3137\n",
      "Epoch 237/2000\n",
      "2/2 - 0s - loss: 0.1873 - val_loss: 0.3137\n",
      "Epoch 238/2000\n",
      "2/2 - 0s - loss: 0.1882 - val_loss: 0.3137\n",
      "Epoch 239/2000\n",
      "2/2 - 0s - loss: 0.1865 - val_loss: 0.3139\n",
      "Epoch 240/2000\n",
      "2/2 - 0s - loss: 0.1848 - val_loss: 0.3138\n",
      "Epoch 241/2000\n",
      "2/2 - 0s - loss: 0.1856 - val_loss: 0.3139\n",
      "Epoch 242/2000\n",
      "2/2 - 0s - loss: 0.1840 - val_loss: 0.3140\n",
      "Epoch 243/2000\n",
      "2/2 - 0s - loss: 0.1823 - val_loss: 0.3140\n",
      "Epoch 244/2000\n",
      "2/2 - 0s - loss: 0.1831 - val_loss: 0.3142\n",
      "Epoch 245/2000\n",
      "2/2 - 0s - loss: 0.1815 - val_loss: 0.3143\n",
      "Epoch 246/2000\n",
      "2/2 - 0s - loss: 0.1797 - val_loss: 0.3141\n",
      "Epoch 247/2000\n",
      "2/2 - 0s - loss: 0.1806 - val_loss: 0.3142\n",
      "Epoch 248/2000\n",
      "2/2 - 0s - loss: 0.1790 - val_loss: 0.3142\n",
      "Epoch 249/2000\n",
      "2/2 - 0s - loss: 0.1776 - val_loss: 0.3142\n",
      "Epoch 250/2000\n",
      "2/2 - 0s - loss: 0.1783 - val_loss: 0.3143\n",
      "Epoch 251/2000\n",
      "2/2 - 0s - loss: 0.1769 - val_loss: 0.3143\n",
      "Epoch 252/2000\n",
      "2/2 - 0s - loss: 0.1756 - val_loss: 0.3142\n",
      "Epoch 253/2000\n",
      "2/2 - 0s - loss: 0.1763 - val_loss: 0.3139\n",
      "Epoch 254/2000\n",
      "2/2 - 0s - loss: 0.1749 - val_loss: 0.3135\n",
      "Epoch 255/2000\n",
      "2/2 - 0s - loss: 0.1736 - val_loss: 0.3137\n",
      "Epoch 256/2000\n",
      "2/2 - 0s - loss: 0.1742 - val_loss: 0.3135\n",
      "Epoch 257/2000\n",
      "2/2 - 0s - loss: 0.1730 - val_loss: 0.3133\n",
      "Epoch 258/2000\n",
      "2/2 - 0s - loss: 0.1716 - val_loss: 0.3134\n",
      "Epoch 259/2000\n",
      "2/2 - 0s - loss: 0.1723 - val_loss: 0.3133\n",
      "Epoch 260/2000\n",
      "2/2 - 0s - loss: 0.1711 - val_loss: 0.3131\n",
      "Epoch 261/2000\n",
      "2/2 - 0s - loss: 0.1698 - val_loss: 0.3131\n",
      "Epoch 262/2000\n",
      "2/2 - 0s - loss: 0.1704 - val_loss: 0.3131\n",
      "Epoch 263/2000\n",
      "2/2 - 0s - loss: 0.1690 - val_loss: 0.3132\n",
      "Epoch 264/2000\n",
      "2/2 - 0s - loss: 0.1677 - val_loss: 0.3133\n",
      "Epoch 265/2000\n",
      "2/2 - 0s - loss: 0.1684 - val_loss: 0.3134\n",
      "Epoch 266/2000\n",
      "2/2 - 0s - loss: 0.1670 - val_loss: 0.3135\n",
      "Epoch 267/2000\n",
      "2/2 - 0s - loss: 0.1658 - val_loss: 0.3134\n",
      "Epoch 268/2000\n",
      "2/2 - 0s - loss: 0.1664 - val_loss: 0.3136\n",
      "Epoch 269/2000\n",
      "2/2 - 0s - loss: 0.1651 - val_loss: 0.3138\n",
      "Epoch 270/2000\n",
      "2/2 - 0s - loss: 0.1640 - val_loss: 0.3137\n",
      "Epoch 271/2000\n",
      "2/2 - 0s - loss: 0.1645 - val_loss: 0.3139\n",
      "Epoch 272/2000\n",
      "2/2 - 0s - loss: 0.1633 - val_loss: 0.3141\n",
      "Epoch 273/2000\n",
      "2/2 - 0s - loss: 0.1621 - val_loss: 0.3138\n",
      "Epoch 274/2000\n",
      "2/2 - 0s - loss: 0.1627 - val_loss: 0.3135\n",
      "Epoch 275/2000\n",
      "2/2 - 0s - loss: 0.1616 - val_loss: 0.3131\n",
      "Epoch 276/2000\n",
      "2/2 - 0s - loss: 0.1606 - val_loss: 0.3132\n",
      "Epoch 277/2000\n",
      "2/2 - 0s - loss: 0.1611 - val_loss: 0.3128\n",
      "Epoch 278/2000\n",
      "2/2 - 0s - loss: 0.1600 - val_loss: 0.3125\n",
      "Epoch 279/2000\n",
      "2/2 - 0s - loss: 0.1589 - val_loss: 0.3127\n",
      "Epoch 280/2000\n",
      "2/2 - 0s - loss: 0.1595 - val_loss: 0.3127\n",
      "Epoch 281/2000\n",
      "2/2 - 0s - loss: 0.1584 - val_loss: 0.3126\n",
      "Epoch 282/2000\n",
      "2/2 - 0s - loss: 0.1573 - val_loss: 0.3126\n",
      "Epoch 283/2000\n",
      "2/2 - 0s - loss: 0.1579 - val_loss: 0.3125\n",
      "Epoch 284/2000\n",
      "2/2 - 0s - loss: 0.1568 - val_loss: 0.3123\n",
      "Epoch 285/2000\n",
      "2/2 - 0s - loss: 0.1558 - val_loss: 0.3125\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 286/2000\n",
      "2/2 - 0s - loss: 0.1563 - val_loss: 0.3125\n",
      "Epoch 287/2000\n",
      "2/2 - 0s - loss: 0.1553 - val_loss: 0.3126\n",
      "Epoch 288/2000\n",
      "2/2 - 0s - loss: 0.1543 - val_loss: 0.3125\n",
      "Epoch 289/2000\n",
      "2/2 - 0s - loss: 0.1548 - val_loss: 0.3125\n",
      "Epoch 290/2000\n",
      "2/2 - 0s - loss: 0.1537 - val_loss: 0.3125\n",
      "Epoch 291/2000\n",
      "2/2 - 0s - loss: 0.1527 - val_loss: 0.3126\n",
      "Epoch 292/2000\n",
      "2/2 - 0s - loss: 0.1532 - val_loss: 0.3127\n",
      "Epoch 293/2000\n",
      "2/2 - 0s - loss: 0.1522 - val_loss: 0.3127\n",
      "Epoch 294/2000\n",
      "2/2 - 0s - loss: 0.1512 - val_loss: 0.3125\n",
      "Epoch 295/2000\n",
      "2/2 - 0s - loss: 0.1517 - val_loss: 0.3124\n",
      "Epoch 296/2000\n",
      "2/2 - 0s - loss: 0.1509 - val_loss: 0.3124\n",
      "Epoch 297/2000\n",
      "2/2 - 0s - loss: 0.1500 - val_loss: 0.3124\n",
      "Epoch 298/2000\n",
      "2/2 - 0s - loss: 0.1504 - val_loss: 0.3124\n",
      "Epoch 299/2000\n",
      "2/2 - 0s - loss: 0.1494 - val_loss: 0.3124\n",
      "Epoch 300/2000\n",
      "2/2 - 0s - loss: 0.1485 - val_loss: 0.3124\n",
      "Epoch 301/2000\n",
      "2/2 - 0s - loss: 0.1490 - val_loss: 0.3123\n",
      "Epoch 302/2000\n",
      "2/2 - 0s - loss: 0.1481 - val_loss: 0.3121\n",
      "Epoch 303/2000\n",
      "2/2 - 0s - loss: 0.1473 - val_loss: 0.3122\n",
      "Epoch 304/2000\n",
      "2/2 - 0s - loss: 0.1477 - val_loss: 0.3121\n",
      "Epoch 305/2000\n",
      "2/2 - 0s - loss: 0.1468 - val_loss: 0.3123\n",
      "Epoch 306/2000\n",
      "2/2 - 0s - loss: 0.1460 - val_loss: 0.3124\n",
      "Epoch 307/2000\n",
      "2/2 - 0s - loss: 0.1463 - val_loss: 0.3127\n",
      "Epoch 308/2000\n",
      "2/2 - 0s - loss: 0.1455 - val_loss: 0.3131\n",
      "Epoch 309/2000\n",
      "2/2 - 0s - loss: 0.1446 - val_loss: 0.3129\n",
      "Epoch 310/2000\n",
      "2/2 - 0s - loss: 0.1450 - val_loss: 0.3133\n",
      "Epoch 311/2000\n",
      "2/2 - 0s - loss: 0.1441 - val_loss: 0.3136\n",
      "Epoch 312/2000\n",
      "2/2 - 0s - loss: 0.1432 - val_loss: 0.3134\n",
      "Epoch 313/2000\n",
      "2/2 - 0s - loss: 0.1437 - val_loss: 0.3136\n",
      "Epoch 314/2000\n",
      "2/2 - 0s - loss: 0.1428 - val_loss: 0.3133\n",
      "Epoch 315/2000\n",
      "2/2 - 0s - loss: 0.1420 - val_loss: 0.3133\n",
      "Epoch 316/2000\n",
      "2/2 - 0s - loss: 0.1424 - val_loss: 0.3133\n",
      "Epoch 317/2000\n",
      "2/2 - 0s - loss: 0.1416 - val_loss: 0.3134\n",
      "Epoch 318/2000\n",
      "2/2 - 0s - loss: 0.1408 - val_loss: 0.3134\n",
      "Epoch 319/2000\n",
      "2/2 - 0s - loss: 0.1412 - val_loss: 0.3133\n",
      "Epoch 320/2000\n",
      "2/2 - 0s - loss: 0.1405 - val_loss: 0.3133\n",
      "Epoch 321/2000\n",
      "2/2 - 0s - loss: 0.1398 - val_loss: 0.3133\n",
      "Epoch 322/2000\n",
      "2/2 - 0s - loss: 0.1400 - val_loss: 0.3133\n",
      "Epoch 323/2000\n",
      "2/2 - 0s - loss: 0.1393 - val_loss: 0.3135\n",
      "Epoch 324/2000\n",
      "2/2 - 0s - loss: 0.1385 - val_loss: 0.3133\n",
      "Epoch 325/2000\n",
      "2/2 - 0s - loss: 0.1389 - val_loss: 0.3131\n",
      "Epoch 326/2000\n",
      "2/2 - 0s - loss: 0.1381 - val_loss: 0.3131\n",
      "Epoch 327/2000\n",
      "2/2 - 0s - loss: 0.1374 - val_loss: 0.3132\n",
      "Epoch 328/2000\n",
      "2/2 - 0s - loss: 0.1378 - val_loss: 0.3132\n",
      "Epoch 329/2000\n",
      "2/2 - 0s - loss: 0.1370 - val_loss: 0.3131\n",
      "Epoch 330/2000\n",
      "2/2 - 0s - loss: 0.1363 - val_loss: 0.3132\n",
      "Epoch 331/2000\n",
      "2/2 - 0s - loss: 0.1367 - val_loss: 0.3131\n",
      "Epoch 332/2000\n",
      "2/2 - 0s - loss: 0.1360 - val_loss: 0.3129\n",
      "Epoch 333/2000\n",
      "2/2 - 0s - loss: 0.1353 - val_loss: 0.3129\n",
      "Epoch 334/2000\n",
      "2/2 - 0s - loss: 0.1356 - val_loss: 0.3126\n",
      "Epoch 335/2000\n",
      "2/2 - 0s - loss: 0.1349 - val_loss: 0.3122\n",
      "Epoch 336/2000\n",
      "2/2 - 0s - loss: 0.1343 - val_loss: 0.3124\n",
      "Epoch 337/2000\n",
      "2/2 - 0s - loss: 0.1346 - val_loss: 0.3123\n",
      "Epoch 338/2000\n",
      "2/2 - 0s - loss: 0.1339 - val_loss: 0.3123\n",
      "Epoch 339/2000\n",
      "2/2 - 0s - loss: 0.1332 - val_loss: 0.3124\n",
      "Epoch 340/2000\n",
      "2/2 - 0s - loss: 0.1335 - val_loss: 0.3125\n",
      "Epoch 341/2000\n",
      "2/2 - 0s - loss: 0.1328 - val_loss: 0.3127\n",
      "Epoch 342/2000\n",
      "2/2 - 0s - loss: 0.1321 - val_loss: 0.3126\n",
      "Epoch 343/2000\n",
      "2/2 - 0s - loss: 0.1325 - val_loss: 0.3126\n",
      "Epoch 344/2000\n",
      "2/2 - 0s - loss: 0.1318 - val_loss: 0.3127\n",
      "Epoch 345/2000\n",
      "2/2 - 0s - loss: 0.1311 - val_loss: 0.3127\n",
      "Epoch 346/2000\n",
      "2/2 - 0s - loss: 0.1315 - val_loss: 0.3128\n",
      "Epoch 347/2000\n",
      "2/2 - 0s - loss: 0.1308 - val_loss: 0.3130\n",
      "Epoch 348/2000\n",
      "2/2 - 0s - loss: 0.1302 - val_loss: 0.3129\n",
      "Epoch 349/2000\n",
      "2/2 - 0s - loss: 0.1305 - val_loss: 0.3130\n",
      "Epoch 350/2000\n",
      "2/2 - 0s - loss: 0.1299 - val_loss: 0.3132\n",
      "Epoch 351/2000\n",
      "2/2 - 0s - loss: 0.1292 - val_loss: 0.3131\n",
      "Epoch 352/2000\n",
      "2/2 - 0s - loss: 0.1296 - val_loss: 0.3133\n",
      "Epoch 353/2000\n",
      "2/2 - 0s - loss: 0.1289 - val_loss: 0.3135\n",
      "Epoch 354/2000\n",
      "2/2 - 0s - loss: 0.1283 - val_loss: 0.3135\n",
      "Epoch 00354: early stopping\n",
      "CPU times: user 1min 57s, sys: 26.2 s, total: 2min 23s\n",
      "Wall time: 20 s\n"
     ]
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 25;\n",
       "                var nbb_unformatted_code = \"%%time\\nhistory = model.fit(\\n    input_train,\\n    y_train.reshape(-1, 1),\\n    epochs=2000,\\n    batch_size=1024,\\n    validation_data=(input_valid, y_valid.reshape(-1, 1),),\\n    verbose=2,\\n    callbacks=[EarlyStopping(monitor=\\\"val_loss\\\", patience=50, verbose=1)],\\n    #class_weight=None#class_weight\\n)\";\n",
       "                var nbb_formatted_code = \"%%time\\nhistory = model.fit(\\n    input_train,\\n    y_train.reshape(-1, 1),\\n    epochs=2000,\\n    batch_size=1024,\\n    validation_data=(input_valid, y_valid.reshape(-1, 1),),\\n    verbose=2,\\n    callbacks=[EarlyStopping(monitor=\\\"val_loss\\\", patience=50, verbose=1)],\\n    #class_weight=None#class_weight\\n)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%time\n",
    "history = model.fit(\n",
    "    input_train,\n",
    "    y_train.reshape(-1, 1),\n",
    "    epochs=2000,\n",
    "    batch_size=1024,\n",
    "    validation_data=(input_valid, y_valid.reshape(-1, 1),),\n",
    "    verbose=2,\n",
    "    callbacks=[EarlyStopping(monitor=\"val_loss\", patience=50, verbose=1)],\n",
    "    #class_weight=None#class_weight\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:34.532826Z",
     "start_time": "2020-09-19T17:12:34.401216Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 26;\n",
       "                var nbb_unformatted_code = \"plot_history(history)\";\n",
       "                var nbb_formatted_code = \"plot_history(history)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_history(history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:34.831294Z",
     "start_time": "2020-09-19T17:12:34.533801Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "21797.53396381579"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 27;\n",
       "                var nbb_unformatted_code = \"model_auc = mean_absolute_error(\\n    y_pred=target_encoder.inverse_transform(model.predict(input_valid)).reshape(-1),\\n    y_true=train[target].values[valid_indices],\\n)\\nmodel_auc\";\n",
       "                var nbb_formatted_code = \"model_auc = mean_absolute_error(\\n    y_pred=target_encoder.inverse_transform(model.predict(input_valid)).reshape(-1),\\n    y_true=train[target].values[valid_indices],\\n)\\nmodel_auc\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_auc = mean_absolute_error(\n",
    "    y_pred=target_encoder.inverse_transform(model.predict(input_valid)).reshape(-1),\n",
    "    y_true=train[target].values[valid_indices],\n",
    ")\n",
    "model_auc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:34.884934Z",
     "start_time": "2020-09-19T17:12:34.832182Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1977252545.9905922"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 28;\n",
       "                var nbb_unformatted_code = \"model_auc = mean_squared_error(\\n    y_pred=target_encoder.inverse_transform(model.predict(input_valid)).reshape(-1),\\n    y_true=train[target].values[valid_indices],\\n)\\nmodel_auc\";\n",
       "                var nbb_formatted_code = \"model_auc = mean_squared_error(\\n    y_pred=target_encoder.inverse_transform(model.predict(input_valid)).reshape(-1),\\n    y_true=train[target].values[valid_indices],\\n)\\nmodel_auc\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_auc = mean_squared_error(\n",
    "    y_pred=target_encoder.inverse_transform(model.predict(input_valid)).reshape(-1),\n",
    "    y_true=train[target].values[valid_indices],\n",
    ")\n",
    "model_auc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:34.938457Z",
     "start_time": "2020-09-19T17:12:34.885867Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "20278.334011130137"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 29;\n",
       "                var nbb_unformatted_code = \"model_auc = mean_absolute_error(\\n    y_pred=target_encoder.inverse_transform(model.predict(input_test)).reshape(-1),\\n    y_true=train[target].values[test_indices],\\n)\\nmodel_auc\";\n",
       "                var nbb_formatted_code = \"model_auc = mean_absolute_error(\\n    y_pred=target_encoder.inverse_transform(model.predict(input_test)).reshape(-1),\\n    y_true=train[target].values[test_indices],\\n)\\nmodel_auc\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_auc = mean_absolute_error(\n",
    "    y_pred=target_encoder.inverse_transform(model.predict(input_test)).reshape(-1),\n",
    "    y_true=train[target].values[test_indices],\n",
    ")\n",
    "model_auc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:34.991577Z",
     "start_time": "2020-09-19T17:12:34.939324Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1481659674.2370112"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 30;\n",
       "                var nbb_unformatted_code = \"model_auc = mean_squared_error(\\n    y_pred=target_encoder.inverse_transform(model.predict(input_test)).reshape(-1),\\n    y_true=train[target].values[test_indices],\\n)\\nmodel_auc\";\n",
       "                var nbb_formatted_code = \"model_auc = mean_squared_error(\\n    y_pred=target_encoder.inverse_transform(model.predict(input_test)).reshape(-1),\\n    y_true=train[target].values[test_indices],\\n)\\nmodel_auc\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_auc = mean_squared_error(\n",
    "    y_pred=target_encoder.inverse_transform(model.predict(input_test)).reshape(-1),\n",
    "    y_true=train[target].values[test_indices],\n",
    ")\n",
    "model_auc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:35.015424Z",
     "start_time": "2020-09-19T17:12:34.992721Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 31;\n",
       "                var nbb_unformatted_code = \"from xplainet.model import predict, encode\";\n",
       "                var nbb_formatted_code = \"from xplainet.model import predict, encode\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from xplainet.model import predict, encode"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:35.302259Z",
     "start_time": "2020-09-19T17:12:35.016264Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 32;\n",
       "                var nbb_unformatted_code = \"probs, explanations = predict(model, input_test)\";\n",
       "                var nbb_formatted_code = \"probs, explanations = predict(model, input_test)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "probs, explanations = predict(model, input_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:35.588774Z",
     "start_time": "2020-09-19T17:12:35.303353Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 33;\n",
       "                var nbb_unformatted_code = \"probs, encoded_output = encode(model, input_test)\";\n",
       "                var nbb_formatted_code = \"probs, encoded_output = encode(model, input_test)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "probs, encoded_output = encode(model, input_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:35.614752Z",
     "start_time": "2020-09-19T17:12:35.589959Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(146, 1)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 34;\n",
       "                var nbb_unformatted_code = \"y_test.shape\";\n",
       "                var nbb_formatted_code = \"y_test.shape\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:35.639749Z",
     "start_time": "2020-09-19T17:12:35.615859Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(146, 384)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 35;\n",
       "                var nbb_unformatted_code = \"encoded_output.shape\";\n",
       "                var nbb_formatted_code = \"encoded_output.shape\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "encoded_output.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:35.666661Z",
     "start_time": "2020-09-19T17:12:35.641100Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(146, 80)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 36;\n",
       "                var nbb_unformatted_code = \"explanations.shape\";\n",
       "                var nbb_formatted_code = \"explanations.shape\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "explanations.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:35.698828Z",
     "start_time": "2020-09-19T17:12:35.667736Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 37;\n",
       "                var nbb_unformatted_code = \"import matplotlib.pyplot as plt\\n\\n# plt.rcdefaults()\\nimport numpy as np\\nimport matplotlib.pyplot as plt\\n\\n\\ndef explain_plot(importances, columns):\\n    selection = np.argsort(-np.absolute(importances))[:10]\\n    # indexes = np.argsort(importances)\\n    performance = importances[selection]\\n    #     print(performance.shape)512, 256, 128, 64, 32, 1\\n    y_pos = np.arange(performance.shape[0])\\n\\n    plt.barh(y_pos, performance, align=\\\"center\\\", alpha=0.5)\\n    plt.yticks(y_pos, columns[selection])\\n    # plt.xlabel('Usage')\\n    plt.title(\\\"Feature importance\\\")\\n\\n    plt.show()\";\n",
       "                var nbb_formatted_code = \"import matplotlib.pyplot as plt\\n\\n# plt.rcdefaults()\\nimport numpy as np\\nimport matplotlib.pyplot as plt\\n\\n\\ndef explain_plot(importances, columns):\\n    selection = np.argsort(-np.absolute(importances))[:10]\\n    # indexes = np.argsort(importances)\\n    performance = importances[selection]\\n    #     print(performance.shape)512, 256, 128, 64, 32, 1\\n    y_pos = np.arange(performance.shape[0])\\n\\n    plt.barh(y_pos, performance, align=\\\"center\\\", alpha=0.5)\\n    plt.yticks(y_pos, columns[selection])\\n    # plt.xlabel('Usage')\\n    plt.title(\\\"Feature importance\\\")\\n\\n    plt.show()\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# plt.rcdefaults()\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "def explain_plot(importances, columns):\n",
    "    selection = np.argsort(-np.absolute(importances))[:10]\n",
    "    # indexes = np.argsort(importances)\n",
    "    performance = importances[selection]\n",
    "    #     print(performance.shape)512, 256, 128, 64, 32, 1\n",
    "    y_pos = np.arange(performance.shape[0])\n",
    "\n",
    "    plt.barh(y_pos, performance, align=\"center\", alpha=0.5)\n",
    "    plt.yticks(y_pos, columns[selection])\n",
    "    # plt.xlabel('Usage')\n",
    "    plt.title(\"Feature importance\")\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:35.726818Z",
     "start_time": "2020-09-19T17:12:35.700042Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Alley', 'Street', 'CentralAir', 'Utilities', 'GarageYrBlt',\n",
       "       '1stFlrSF', 'LotFrontage', 'GrLivArea', 'ScreenPorch',\n",
       "       'TotalBsmtSF', 'BsmtFinSF1', 'Id', 'BsmtUnfSF', 'GarageArea',\n",
       "       'YearBuilt', 'BsmtFinSF2', 'WoodDeckSF', 'MasVnrArea', 'LotArea',\n",
       "       'EnclosedPorch', '2ndFlrSF', 'YearRemodAdd', 'OpenPorchSF',\n",
       "       'HeatingQC', 'Condition2', 'MiscFeature', 'RoofStyle', 'BsmtCond',\n",
       "       'Foundation', 'PavedDrive', 'LandSlope', 'MiscVal', 'LotConfig',\n",
       "       'GarageCond', 'ExterCond', 'SaleType', 'Fence', 'KitchenQual',\n",
       "       'SaleCondition', 'Fireplaces', '3SsnPorch', 'GarageFinish',\n",
       "       'PoolQC', 'RoofMatl', 'BsmtHalfBath', 'GarageType', 'Heating',\n",
       "       'BsmtFullBath', 'Electrical', 'MSSubClass', 'Exterior1st',\n",
       "       'ExterQual', 'LandContour', 'TotRmsAbvGrd', 'LotShape',\n",
       "       'BsmtExposure', 'BedroomAbvGr', 'OverallQual', 'OverallCond',\n",
       "       'BsmtFinType2', 'Neighborhood', 'FullBath', 'Functional',\n",
       "       'FireplaceQu', 'MasVnrType', 'HalfBath', 'MoSold', 'YrSold',\n",
       "       'BsmtQual', 'BsmtFinType1', 'GarageCars', 'BldgType', 'MSZoning',\n",
       "       'HouseStyle', 'Exterior2nd', 'KitchenAbvGr', 'LowQualFinSF',\n",
       "       'Condition1', 'PoolArea', 'GarageQual'], dtype='<U13')"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 38;\n",
       "                var nbb_unformatted_code = \"all_cols = np.array(params[\\\"bool_cols\\\"] + params[\\\"num_cols\\\"] + params[\\\"cat_cols\\\"])\\nall_cols\";\n",
       "                var nbb_formatted_code = \"all_cols = np.array(params[\\\"bool_cols\\\"] + params[\\\"num_cols\\\"] + params[\\\"cat_cols\\\"])\\nall_cols\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "all_cols = np.array(params[\"bool_cols\"] + params[\"num_cols\"] + params[\"cat_cols\"])\n",
    "all_cols"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-18T18:48:54.151671Z",
     "start_time": "2020-09-18T18:48:54.127488Z"
    }
   },
   "source": [
    "## Explain global"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:35.751197Z",
     "start_time": "2020-09-19T17:12:35.727963Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 39;\n",
       "                var nbb_unformatted_code = \"# Looks like this is broken for now.\";\n",
       "                var nbb_formatted_code = \"# Looks like this is broken for now.\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Looks like this is broken for now."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:36.068470Z",
     "start_time": "2020-09-19T17:12:35.752238Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 40;\n",
       "                var nbb_unformatted_code = \"probs_train, explanations_train = predict(model, input_train)\";\n",
       "                var nbb_formatted_code = \"probs_train, explanations_train = predict(model, input_train)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "probs_train, explanations_train = predict(model, input_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:36.093578Z",
     "start_time": "2020-09-19T17:12:36.069411Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 41;\n",
       "                var nbb_unformatted_code = \"global_explain = np.abs(explanations_train).sum(axis=0)\\nglobal_explain = global_explain / np.abs(global_explain).sum()\";\n",
       "                var nbb_formatted_code = \"global_explain = np.abs(explanations_train).sum(axis=0)\\nglobal_explain = global_explain / np.abs(global_explain).sum()\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "global_explain = np.abs(explanations_train).sum(axis=0)\n",
    "global_explain = global_explain / np.abs(global_explain).sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:36.117950Z",
     "start_time": "2020-09-19T17:12:36.094608Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.01215346, 0.01259931, 0.01232962, 0.01247911, 0.01251009,\n",
       "       0.01259026, 0.01250862, 0.01258593, 0.01249374, 0.01257558,\n",
       "       0.01252444, 0.01247921, 0.0125008 , 0.0125607 , 0.0124805 ,\n",
       "       0.01247848, 0.01250196, 0.01253927, 0.01253721, 0.01247924,\n",
       "       0.01254214, 0.01256049, 0.01251706, 0.01246301, 0.01246805,\n",
       "       0.01246541, 0.01255347, 0.0124383 , 0.01249522, 0.01243638,\n",
       "       0.01254407, 0.0124729 , 0.01250988, 0.01246   , 0.01247715,\n",
       "       0.01254957, 0.01247233, 0.01256865, 0.0124852 , 0.01253109,\n",
       "       0.01250283, 0.01250565, 0.01249047, 0.01249789, 0.01244909,\n",
       "       0.01245513, 0.01248395, 0.01252677, 0.01246678, 0.01246023,\n",
       "       0.0125155 , 0.01259258, 0.01251211, 0.01258737, 0.01252522,\n",
       "       0.0126105 , 0.01253636, 0.01262202, 0.01245211, 0.01246091,\n",
       "       0.0126368 , 0.01253023, 0.01242545, 0.01258366, 0.01250471,\n",
       "       0.0125318 , 0.01246474, 0.01247848, 0.01255076, 0.01250262,\n",
       "       0.0126519 , 0.0123129 , 0.01240089, 0.01247945, 0.01248726,\n",
       "       0.01242152, 0.01247916, 0.01240441, 0.01250433, 0.01247756])"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 42;\n",
       "                var nbb_unformatted_code = \"global_explain\";\n",
       "                var nbb_formatted_code = \"global_explain\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "global_explain"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:36.247721Z",
     "start_time": "2020-09-19T17:12:36.118856Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 43;\n",
       "                var nbb_unformatted_code = \"explain_plot(global_explain, all_cols)\";\n",
       "                var nbb_formatted_code = \"explain_plot(global_explain, all_cols)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "explain_plot(global_explain, all_cols)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Explain local"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:36.270621Z",
     "start_time": "2020-09-19T17:12:36.248688Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 44;\n",
       "                var nbb_unformatted_code = \"# Looks like this is broken for now.\";\n",
       "                var nbb_formatted_code = \"# Looks like this is broken for now.\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Looks like this is broken for now."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:38.493025Z",
     "start_time": "2020-09-19T17:12:36.271469Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.6638028621673584\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-1.6289162635803223\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.6235273480415344\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.7696265578269958\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbcAAAEICAYAAADcC7pLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAoUUlEQVR4nO3debxVdb3/8dfbERXFATInPGWCsyg7y3lIvDaqqSFZhlZeyzQhLG/1u1qW1fWa5cVukSmmpl7FgbJUHBA1FQ7IoOBQigqSopg5ksDn98f6nlgcz7DPPvvsvc/i/Xw89uOs9V3f8Yj7c77ftfb+KiIwMzMrkjXq3QEzM7Nqc3AzM7PCcXAzM7PCcXAzM7PCcXAzM7PCcXAzM7PCcXAzKzBJ35Z0Sb37YVZr8ufczNomaT6wObA8lzwoIp7vZp1fiog7ute73kfSOcAHIuJz9e6LFZ9nbmYd+2RE9M29Kg5s1SBprXq2X6ne2m/rvRzczLpIUj9Jv5G0SNJCST+QtGa6tp2kuyS9LOklSVdJ2jhduwIYCPxe0uuSvinpIEkLWtU/X9Kh6fgcSddLulLSP4CRHbXfRl/PkXRlOm6SFJJOlPScpFcknSLpg5JmS/q7pLG5siMl3S9prKRXJT0m6SO561tKmihpiaS/SPpyq3bz/T4F+DYwPI19Vsp3oqR5kl6T9JSkf8/VcZCkBZK+IenFNN4Tc9fXk3SBpGdS/+6TtF669mFJf05jmiXpoAr+U1sv5uBm1nXjgWXAB4A9gMOAL6VrAn4EbAnsCGwDnAMQEZ8HnmXlbPC/ymzvCOB6YGPgqk7aL8eHgO2B4cDPgO8AhwI7A5+RdGCrvH8F+gNnAzdI2jRduwZYkMZ6DHCepEPa6fdvgPOAa9PYd095XgQ+AWwEnAhcKGnPXB3vBfoBWwFfBC6WtEm69t/AUGAfYFPgm8AKSVsBtwA/SOljgAmSBnThd2S9nIObWcduSn/9/13STZI2Bz4GnBERb0TEi8CFwHEAEfGXiJgUEUsjYjHwU+DA9qsvywMRcVNErCALAu22X6ZzI+LtiLgdeAO4OiJejIiFwL1kAbPFi8DPIuKdiLgWeBz4uKRtgH2Bb6W6ZgKXACe01e+IeKutjkTELRHx18jcA9wO7J/L8g7w/dT+H4HXgcGS1gBOAr4eEQsjYnlE/DkilgKfA/4YEX9MbU8CmtPvzVYTXgc369iR+Yc/JO0FrA0sktSSvAbwXLq+OfBzsjfoDdO1V7rZh+dyx9t21H6ZXsgdv9XGed/c+cJY9amzZ8hmalsCSyLitVbXSu30u02SPko2IxxENo71gTm5LC9HxLLc+Zupf/2BPmSzyta2BY6V9Mlc2trA3Z31x4rDwc2sa54DlgL9W73ptjgPCGDXiFgi6UhgbO5668eT3yB7Qwcg3TtrvXyWL9NZ+9W2lSTlAtxAYCLwPLCppA1zAW4gsDBXtvVYVzmXtC4wgWy2d3NEvCPpJrKl3c68BLwNbAfManXtOeCKiPjyu0rZasPLkmZdEBGLyJbOLpC0kaQ10kMkLUuPG5Itnb2a7v2c2aqKF4D3586fAPpI+riktYHvAut2o/1qew9wuqS1JR1Ldh/xjxHxHPBn4EeS+kjajeye2JUd1PUC0JSWFAHWIRvrYmBZmsUdVk6n0hLtpcBP04Mta0raOwXMK4FPSvq3lN4nPZyyddeHb72Vg5tZ151A9sY8l2zJ8Xpgi3Tte8CewKtkDzXc0Krsj4Dvpnt4YyLiVeCrZPerFpLN5BbQsY7ar7aHyB4+eQn4IXBMRLycro0AmshmcTcCZ3fy+b3r0s+XJc1IM77Tgf8jG8dnyWaF5RpDtoQ5DVgC/ARYIwXeI8iezlxMNpM7E7/frVb8IW4za5OkkWQfON+v3n0x6yr/JWNmZoXj4GZmZoXjZUkzMyscz9zMzKxw/Dm3Gujfv380NTXVuxtmZr3K9OnTX4qIir42zcGtBpqammhubq53N8zMehVJz1Ra1suSZmZWOA5uZmZWOA5uZmZWOA5uZmZWOA5uZmZWOA5uZmZWOA5uZmZWOA5uZmZWOP4Qdy904aQn6t0FM7OG5pmbmZkVjoObmZkVjoObmZkVjoObmZkVTq8KbpIulfSipEc6yXeQpH1y5+dIWihpZnr9OKVPllRqp45PSHpY0ixJcyX9e0d1mZlZ4+htT0uOB8YCv+0k30HA68Cfc2kXRsR/l9OIpHWBccBeEbEgnTdVUpeZmdVer5q5RcQUYEk+TdLpaWY1W9I1kpqAU4BRaWa1fzl1S3pd0gWSZgEfIgv8L6d2l0bE41UdjJmZ9ZheFdzacRawR0TsBpwSEfOBX5LNroZExL0p36jcUuK/tVHPBsBDEbF7CqITgWckXS3peEn531VndSHpZEnNkpoXL15ctcGamVnnihDcZgNXSfocsKyDfC3BbkhE3NbG9eXAhJaTiPgS8BFgKjAGuLQLdRER4yKiFBGlAQMq2iXdzMwqVITg9nHgYmBPYJqkSu8jvh0Ry/MJETEnIi4EhgFHd6+bZmZWK706uKWlwm0i4m7gW0A/oC/wGrBhN+rtK+mgXNIQ4JmKO2pmZjXVq56WlHQ12ZOQ/SUtAM4FPi+pHyDgooj4u6TfA9dLOgI4rZKmgG9K+hXwFvAGMLIKQzAzsxroVcEtIka0kfyrNvI9AeyWS7q3dZ6U76Dccd/c8WvAx9opc055vTUzs3rp1cuSZmZmbXFwMzOzwulVy5KWGTVsUL27YGbW40Z3o6xnbmZmVjgObmZmVjgObmZmVji+59YLXTjpiXp3wcysoXnmZmZmhePgZmZmhePgZmZmhePgZmZmhdMwwU3S8twGoDMlndVJ/m9X2M46kn4m6S/p9QdJAyvrNUg6R9KYSsubmVn1NdLTkm9FxJAu5P82cF5XGpC0ZiqzITA4IpZLOhG4WdLQiFjRlfrMzKwxNczMrS2S+kl6XNLgdH61pC9L+jGwXprhXZWufU7S1JT2qxTIkPS6pAskzQL2BU4ERrVsTBoRlwGvA4dKapL0SK79MZLOScdfljRN0ixJEyStX8NfhZmZdUEjBbeWYNXyGh4RrwJfA8ZLOg7YJCJ+HRFnkWZ6EXG8pB2B4cC+afa3HDg+1bsB8FBE7A78HXg2Iv7Rqu1mYKdO+ndDRHww1TMP+GJHmSWdLKlZUvPixYvL/iWYmVn3NfyyZERMknQscDGweztlPwIMBaZJAlgPeDFdWw5MqEL/dpH0A2Bjst2+b+soc0SMA8YBlEqlqEL7ZmZWpkYKbm2StAawI/AmsAmwoK1swOUR8R9tXHu7ZQkS+CswUNKGaUPSFkPJAuAyVp3N9skdjweOjIhZkkaS7QhuZmYNqJGWJdszimwZ8LPAZZLWTunv5I7vBI6R9B4ASZtK2rZ1RRHxBnA58NPcPbkTgLeB+4EXgPdI2kzSusAncsU3BBalNo/HzMwaViPN3NaTNDN3fitwGfAlYK+IeE3SFOC7wNlkS36zJc1I992+C9yeZnrvAKcCz7TRzn8A5wOPS1oPWAzsHRFBFjC/D0wFFgKP5cr9P+ChlP8hsmBnZmYNSNl7+upJ0nuBPwH/m+6R9YhSqRTNzc1Vq89fnGxmq4PRhw2eHhGlSso20syt5iLib8Ae9e6HmZlVV2+452ZmZtYlDm5mZlY4q/WyZG81atigenfBzKzHje5GWc/czMyscBzczMyscBzczMyscHzPrZfxZ9zMzDrnmZuZmRWOg5uZmRWOg5uZmRVOWcEtfUt+yyaif5O0MHe+Tqu8Z+R3qZY0X9IcSbMl3dPWt/VXKrX1tqR+ubSRksZ2sZ7tJf1B0l8lTZd0t6QDyiw7X1L/rvbdzMx6TlnBLSJeTrteDwF+CVzYch4R/2yV/Qxg/VZpB0fEbsBksm/1r5YRwDTg05VWIKkPcAswLiK2i4ihwGnA+9vI6wdwzMx6gYqXJSV9RNLDaVZ2qaR1JZ0ObAncLenuNoo9AGyVyjdJekzSeElPSLpK0qGS7pf0pKS9Ur4Dc7PEhyVtmNK3I9sR+7tkQS5vG0mTUz1np/w/lnRqrv/nSBpDtjfbAxExseVaRDwSEeNz+a6QdD9wRZrF3i7pUUmXkG2UamZmDaTS4NaHbGfq4RGxK9lHCr4SERcBz5PN1A5uo9zhwE258w8AFwA7pNdngf2AMcC3U54xwKlp1rg/8FZKPw64BrgXGCxp81y9ewFHA7sBx0oqAdcCn8nl+UxK2xmY0cl4dwIOjYgRZHvJ3RcROwM3AgM7KWtmZjVWaXBbE3g6Ilo+dHU50NE9qrslLQQ+ClydS386IuZExArgUeDOtGnoHKAp5bmfbOfs04GNI2JZSh8BXJPKTgCOzdU7KS2lvgXcAOwXEQ+T7bK9paTdgVci4rnWHZV0o6RHJN2QS56Y6iKN80qAiLgFeKWtAUs6WVKzpObFixd38KsxM7Nqq9XTkgcD2wIzge/l0pfmjlfkzleQPmAeET8m2417PeB+STtI2hXYHpgkaT7ZLC6/NNl6B9aW8+uAY4DhZLM2yILqnv/KGHEUMBLYNFf+jbJGmW8wYlxElCKiNGDAgK4WNzOzbqg0uC0HmiR9IJ1/HrgnHb8GbNi6QJpxnQGcIGnT1tfbI2m7NLv7CdnDIzuQBbJzIqIpvbYEtsw9iTlM0qaS1gOOJJv9QRbQjiMLcNeltN8B+0r6VK7Z1g/E5E0hWz5F0keBTcodi5mZ1Ualwe1t4ETgOklzyGZav0zXxgG3tvVASUQsIluWPLX1tQ6ckZYJZwPvAH8iC1A3tsp3Y0oHmEq2VDkbmBARzan9R8kC78LUF9Jy4yeAUyQ9JekBsodUftBOf74HHCDpUbKnNJ/twljMzKwGlN3isp5UKpWiubm5KnX5uyXNbHUx+rDB0yOiVElZf0OJmZkVjoObmZkVjoObmZkVjr9OqpcZNWxQvbtgZlYTo7tR1jM3MzMrHAc3MzMrHAc3MzMrHN9z62X8OTczs8555mZmZoXj4GZmZoXj4GZmZoXj4GZmZoXT48FN0nJJM3OvJkl/rmL98yX1r1Z9qc79JE2V9JikxyV9tZr1m5lZz6rF05JvRcSQVmn7tM4kaa3cLtt1I+m9ZHu8HRkRM1LgvE3Soohovc2OmZk1oLosS0p6Pf08SNK9kiYCcyWtKel8SdMkzZb077l8UyTdkmZSv5T0rr5LuknSdEmPSjo5l364pBmSZkm6M6VtIOnSNEN7WNIRKfupwPiImAEQES8B3wTOTOXGSzqm9VjMzKxx1GLmtp6kmen46Yg4qtX1PYFdIuLpFJBejYgPSloXuF/S7SnfXsBOwDPArWQbhV7fqq6TImJJ2oF7mqQJZAH818ABqY2WXcC/A9wVESdJ2hiYKukOYGfg8lb1Nqe2y5bGcjLAwIEDu1LUzMy6qV7LknlTI+LpdHwYsFtuZtQP2B74Z8r3FICkq4H9eHdwO11SS/DcJpUdAExpaSMiluTa+pSkMem8D1C1KBQR48h2JadUKnlHWDOzGmqEbyh5I3cs4LSIuC2fQdJBQOsAEW3kORTYOyLelDSZLGC1R8DREfF4q3rmAkOBm3PJQ8lmbwDLSMu5aWl0nQ7aMDOzOmi0jwLcBnxF0toAkgZJ2iBd20vS+1JAGQ7c16psP+CVFNh2AD6c0h8EDpD0vlRny7LkbcBpkpTS90jpFwMjJQ1J6ZsBPwTOTdfnkwU7gE8Ba3d71GZmVlWNFtwuAeYCMyQ9AvyKlbPLacBYYB7wNND6ycVbgbUkzQN+TBbUiIjFZPe+bpA0C7g25T+XLDDNlvRoOiciFgGfA8ZJehx4HrgoIu5J5X4NHJjq2ptVZ55mZtYAFNH4t4PSkuOYiPhEHdr+KvAVsgdSXqmkjlKpFM3NzZ1nLIO/ONnMVhejDxs8PSJKlZRttJlbw4mIX0TErpUGNjMzq71GeKCkUxExGZhc526YmVkv0SuCm600atigenfBzKwmRnejrJclzcyscBzczMyscBzczMyscHzPrZfxRwHMzDrnmZuZmRWOg5uZmRWOg5uZmRWOg5uZmRVOTYKbpK0l3SzpSUl/lfRzST26VUxut++m9CXMLen7pd23H0u7en+1Gu2YmVnj6PHglraUuQG4KSK2BwYBfcm2kelOvV1+0lPSe4HfAadExA7AvsAXcxucmplZAdRi5nYI8HZEXAYQEcuBUcBJaQa1c0tGSZMllSRtIOnSdP1hSUek6yMlTZR0F3CnpL6S7pQ0Q9KclnwdOBUYHxEzUl9eAr4JnJnqH5/bBTw/++tqO2ZmVke1+JzbzsD0fEJE/EPSs8AtwGeAsyVtAWwREc2SzgPuioiTJG0MTJV0Ryq+J7BbRCxJs7ejUn39gQclTYz29/HZGbi8VVozsFMnY3i7i+0g6WSyfeQYOHBgJ9WbmVk11fuBkslAy0zpM8D16fgw4CxJM1OePkBLhJgUEUvSsYDzJM0G7gC2AjbvgX52uZ2IGBcRpYgoDRgwoAe6ZGZm7anFzG0uKwMYAJI2IgtW04CXJe0GDAdOackCHB0Rj7cq9yFW3fn6eGAAMDQi3pE0nywQdtSXocDNubShZLM3gGWkgC9pDaDloZeutmNmZnVUi5nbncD6kk4AkLQmcAHZva83gWvJ7nv1i4jZqcxtwGnpYRQk7dFO3f2AF1PAORjYtpO+XAyMlDQk1bsZ2YMt56br88mCHcCngLUrbMfMzOqox4Nbui91FHCspCeBJ8juYX07ZbkeOA74v1yxc8kCy2xJj7Iy+LR2FVCSNAc4AXisk74sAj4HjJP0OPA8cFFE3JOy/Bo4UNIsYG9WzhK71I6ZmdWXOngmovDSZ9y+AhwQEa/0VDulUimam5s7z1gGf3Gyma0uRh82eHpElCopW+8HSuoqIn4REbv2ZGAzM7PaW62Dm5mZFZP3c+tlRg0bVO8umJnVxOhulPXMzczMCsfBzczMCsfBzczMCsf33HoZfxTAzKxznrmZmVnhOLiZmVnhOLiZmVnhOLiZmVnhdDm4SVouaaakWWln6n262wlJQyR9LHc+UtLi1E7Lq7MNRc3MzIDKnpZ8KyKGAEj6N+BHwIHd7McQoAT8MZd2bUR8rZv1Vp2ktSJiWb37YWZm7evusuRGwCsAkraQNCXNsh6RtH9Kf13S+ZIelXSHpL0kTZb0lKRPSVoH+D4wPJUd3l5jko6SdKcyW0h6QtJ700zv5lTvk5LOzpUZnfrziKQzUtoGkm5Js89HWtqUNF9S/3RckjQ5HZ8j6QpJ9wNXSBogaYKkaem1bzd/j2ZmVkWVzNzWkzSTbCfqLYBDUvpngdsi4odpQ9L1U/oGwF0RcaakG4EfAMOAnYDLI2KipP8ESi0zNUkjyYLdfrl2946IGyUdDZwKHA6cHRF/S3ua7gXsArwJTJN0CxDAicCHyHb3fkjSPcD7gecj4uOpvX5ljHsnYL+IeEvS74ALI+I+SQPJNlfdMZ9Z0snAyQADBw4so3ozM6uW7i5L7g38VtIuwDTgUklrAzdFxMyU/5/Arel4DrA07Wg9B2jqoJ32liVPAx4BHoyIq3PpkyLi5dSvG4D9yILbjRHxRi59/9SfCyT9BPhDRNxbxrgnRsRb6fhQYKcUVAE2ktQ3Il5vSYiIccA4yPZzK6N+MzOrkm4tS0bEA0B/YEBETAEOABYC4yWdkLK9Eyt3RF0BLE1lV1BZcN061bO5pHz/WweQdgNKRDwB7EkWbH+QZo4Ay1j5O+nTqtgbueM1gA9HxJD02iof2MzMrL66Fdwk7QCsCbwsaVvghYj4NXAJWfAo12vAhmW0txZwKTACmMeqOyIMk7SppPWAI4H7gXuBIyWtL2kD4CjgXklbAm9GxJXA+bm+zgeGpuOjO+jK7WQzyJZ+Dems72ZmVjvduecG2X2sL0TEckkHAWdKegd4HTih7eJtuhs4K9X7o5TW+p7bV8mWA+9N97pmsfLeGsBUYALZzO7KiGgGkDQ+XQO4JCIeTk95ni9pBfAO8JV0/XvAbySdC0zuoL+nAxdLmk32O5wCnNKF8ZqZWQ/SyhXD3is9gFJq5x5d3ZVKpWhubq5KXf7iZDNbXYw+bPD0iChVUtbfUGJmZoVTiC1vImI8ML7O3TAzswZRiOC2Ohk1bFC9u2BmVhOjO8/SLi9LmplZ4Ti4mZlZ4Ti4mZlZ4fieWy/ijwGYmZXHMzczMyscBzczMyscBzczMyscBzczMyucDoNb2vH6PkkfzaUdK+nWjsp1RtLy3I7dv5e0cXfq62LbIyWNbZU2U9I1HZQ5SNIf2rn2r927zcysMXQY3NI+bKcAP5XUR1Jf4DyynbC7LG1ZA2nD04jYBVhSaX3VIGlHsm179k/b4piZWS/X6bJkRDwC/B74FvCfwJXAdyRNlfSwpCMAJDVJulfSjPTaJ6UflNInAnPbaOIBYKuUdztJt0qansrskNLHS/pfSQ9KeirVeamkeWlLG1K+EZLmpBnhT3LpJ0p6QtJUYN9W7Y8AriDbo+2IXJnDJT0maQbw6Vz6ZpJul/SopEvItv0xM7MGUu49t+8BnwU+SrZD9V0RsRdwMNm+aBsALwLDImJPYDhwUa78nsDXI2KVL0aUtCbwEWBiShoHnBYRQ4ExwC9y2TcB9gZGpfwXAjsDu0oakjYg/QlwCDAE+KCkIyVtkfq/L7AfsFOrsQ0HrgGuJgt0SOoD/Br4JNnmpe/N5T8buC8idgZuBAa29QuTdLKkZknNixcvbiuLmZn1kLI+xB0Rb0i6lmwT0s8An5Q0Jl3uQ/YG/zwwNu1KvRzIB7KpEfF07rxlw9OtyHbUnpSWPPcBrpP+NRlaN1fm9xERkuaQ7fg9B0DSo0ATsC0wOSIWp/SrgANS2Xz6tS19k1QCXoqIZyUtBC6VtGkaz9MR8WTKdyVwcqrrANJMLiJukfRKO7+zcWTBmlKp1Ps3zTMz60W68g0lK9JLwNER8Xj+oqRzgBeA3clmhG/nLr/Rqq63ImKIpPWB28juuY0H/h4RQ9ppf2muH0tz6SvSON7pwlhajAB2kDQ/nW8EHA1Mq6AuMzNrEJV8FOA24DSl6ZWkPVJ6P2BRRKwAPk/2kEaHIuJN4HTgG8CbwNOSjk31StLuXejXVOBASf3TcucI4B7goZS+maS1gZb61yCbhe4aEU0R0UR2z20E8BjQJGm7VPeIXDtTyJZoSU+RbtKFPpqZWQ1UEtzOBdYGZqclwXNT+i+AL0iaBezAu2drbYqIh4HZZAHkeOCLqY5HyT3gUUY9i4CzgLuBWcD0iLg5pZ9D9uDK/WTLoAD7Awsj4vlcNVPI7sltQrYMeUt6oOTFXJ7vAQeksX8aeLbcPpqZWW0oe9rfelKpVIrm5uZu1+MvTjaz1cnowwZPj4hSJWX9DSVmZlY4Dm5mZlY43s+tFxk1bFDnmczMCmJ0N8p65mZmZoXj4GZmZoXj4GZmZoXje269hD8GYGZWPs/czMyscBzczMyscBzczMyscBzczMyscBo6uEnaRtLdkuamna+/3sXyk9OebUian3bpnple+6Tdwx9pp+waki5Ku3rPkTRN0vvaq6v7ozUzs2pp9KcllwHfiIgZkjYEpkuaFBFzK6zv4Ih4qeVEUlNbmSStRbY1zpbAbhGxQtLWrLrTwSp1mZlZ42jo4Ja2q1mUjl+TNA/YStIvyPZpOxjYGPhiRNwraT3gMrINUx8D1iu3LUkjybaw6Uu2F93NrNyfjohYUKVhmZlZD2vo4JaXZll7kAU1gLUiYi9JHwPOBg4FvgK8GRE7StoNmNGqmrslLQeWRsSH2mhmT7KZ2pI0U7tP0v7AncCVae+5cusyM7M66RXBTVJfYAJwRkT8I20CfkO6PB1oSscHABcBRMRsSbNbVdXZUuKkiFiSyi+QNBg4JL3ulHRsRNxZTl2STibb8JSBAweWN1AzM6uKhn6gBEDS2mSB7aqIuCF3aWn6uZzqBelVdg+PiKUR8aeIOBM4Dziy3IoiYlxElCKiNGDAgCp1z8zMytHQwU3ZFO03wLyI+GkZRaYAn01ldwF260bbe0raMh2vkep6ptL6zMysdhp9WXJf4PPAHEkzU9q3O8j/v8Bl6cGTeWRLlpV6D/BrSeum86nA2G7UZ2ZmNdLQwS0i7gPUxqU/5vK8RLrnFhFvAce1U1dTG2nzgV3S8XhgfO7arcCt5dZlZmaNo6GXJc3MzCrh4GZmZoXj4GZmZoXT0PfcbKVRwwbVuwtmZjU1uhtlPXMzM7PCcXAzM7PCcXAzM7PC8T23XuLCSU/UuwtmZr2GZ25mZlY4Dm5mZlY4Dm5mZlY4Dm5mZlY4DRXcJG0u6XeSnpI0XdIDko5qI1+TpEfaSP++pEPLaGeIpJB0eLX6bmZmjaNhglvau+0mYEpEvD8ihpJ9w//WrfK1+4RnRPxnRNxRRnMjgPvSzzb7kvZwMzOzXqiR3sAPAf4ZEb9sSYiIZyLifySNlDRR0l3Ane1VIGm8pGMkHS7pulz6QZL+kI4FHAuMBIZJ6pPSmyQ9Lum3wCPANpLOlDRN0mxJ38vVd1OaWT4q6eTq/hrMzKy7Gim47QzM6OD6nsAxEXFgGXXdAXxI0gbpfDhwTTreB3g6Iv4KTAY+niu3PfCLiNgZGJzO9wKGAEMlHZDynZRmliXgdEmbte6ApJMlNUtqXrx4cRldNjOzammk4LYKSRdLmiVpWkqaFBFLyikbEcvINhr9ZFrG/Dhwc7o8gpWB7hpWXZp8JiIeTMeHpdfDZEF3B7JgB1lAmwU8CGyTS8/3YVxElCKiNGDAgHK6bWZmVdJI31DyKHB0y0lEnCqpP9Cckt7oYn3XAF8DlgDNEfGapDVTG0dI+g7ZLt+bSdqwjTYE/CgifpWvVNJBwKHA3hHxpqTJQJ8u9s3MzHpQI83c7gL6SPpKLm39btR3D9lS5pdZOVP7CDA7IraJiKaI2BaYALzriUzgNuAkSX0BJG0l6T1AP+CVFNh2AD7cjT6amVkPaJjgFhEBHAkcKOlpSVOBy4FvtVNksKQFudexrepbDvwB+Gj6CdkS5I2t6plAG09NRsTtwO+AByTNAa4HNiRb7lxL0jzgx2RLk2Zm1kCUxRTrSaVSKZqbmzvP2AF/cbKZrW5GHzZ4ekSUKinbMDM3MzOzanFwMzOzwmmkpyWtA6OGDap3F8zMamp0N8p65mZmZoXj4GZmZoXj4GZmZoXje269gD8GYGbWNZ65mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4VQU3CSFpAty52MkndNJmU9JOquTPP/aMbuNa/PTFjgVadmlu9Lyta7XzMwqV+nMbSnw6a4Em4iYGBE/rrC9bkkblpqZ2Wqi0uC2DBgHjGp9QdIASRMkTUuvfVP6SElj0/F2kh6UNEfSDyS9nquir6TrJT0m6SpJyl37ZiozVdIHUl1Nku6SNFvSnZIGpvTxkn4p6SHgv1L5AyT9WdJTLbMtZc6X9Eiqe3gZ6WMlPS7pDuA9Ff4Ozcysh3TnntvFwPGS+rVK/zlwYUR8kGzX60vaKPtz4OcRsSuwoNW1PYAzgJ2A9wP75q69msqMBX6W0v4HuDwidgOuAi7K5d8a2CciWr6ibAtgP+ATZHuxAXwaGALsTrbD9vmStugg/ShgcOrfCcA+bYwPSSdLapbUvHjx4raymJlZD6k4uEXEP4DfAqe3unQoMFbSTGAisFHLbtY5ewPXpePftbo2NSIWRMQKYCbQlLt2de7n3rm6Wuq4gix4tbgubVra4qaIWBERc4HNU9p+wNURsTwiXiDbwfuDHaQfkEt/nmwH8XeJiHERUYqI0oABA9rKYmZmPaS796J+BswALsulrQF8OCLezmdcdXWxQ0tzx8tZtY/RznF73uig7rI7ZGZmvUu3PgoQEUuA/wO+mEu+HTit5UTSkDaKPki2ZAlwXBeaHJ77+UA6/nOujuOBe7tQHyn/cElrShpANjOb2kH6lFz6FsDBXWzPzMx6WDWeIrwA+Fru/HTgYkmzU/1TgFNalTkDuFLSd4BbgVfLbGuTVO9SYERKOw24TNKZwGLgxC72/0aypc1ZZLPBb0bE3yR1lH4IMBd4lpVB1szMGoQiylndq3Kj0vrAWxERko4DRkTEETXvSI2USqVobm6uuLy/ONnMVkejDxs8PSJKlZSt1+e/hpI9dCLg78BJdeqHmZkVUF2CW0TcS/aIvZmZWdX5mzt6gVHDBtW7C2ZmNTe68yzt8hcnm5lZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4dTli5NXN5IWA890kq0/8FINulNrRR0XFHdsHlfvU9SxDY6IDSsp6K/fqoGI6HQrbknNlX77dSMr6riguGPzuHqfoo5NUsXbqXhZ0szMCsfBzczMCsfBrXGMq3cHekhRxwXFHZvH1fsUdWwVj8sPlJiZWeF45mZmZoXj4GZmZoXj4FYnko6V9KikFZLafIRX0jaS7pY0N+X9eq372VXljCvlO1zS45L+IumsWvaxUpI2lTRJ0pPp5ybt5Puv9DuYJ+kiSap1X7uiC+MaKOn2NK65kppq3NUuKXdcKe9GkhZIGlvLPlaqnLFJGiLpgfRvcbak4fXoazk6ez+QtK6ka9P1h8r5t+fgVj+PAJ8GpnSQZxnwjYjYCfgwcKqknWrRuW7odFyS1gQuBj4K7ASM6AXjAjgLuDMitgfuTOerkLQPsC+wG7AL8EHgwFp2sgKdjiv5LXB+ROwI7AW8WKP+VarccQGcS8f/Lzaacsb2JnBCROwMHA78TNLGtetiecp8P/gi8EpEfAC4EPhJZ/U6uNVJRMyLiMc7ybMoImak49eAecBWtehfpcoZF9kb418i4qmI+CdwDXBEz/eu244ALk/HlwNHtpEngD7AOsC6wNrAC7XoXDd0Oq70ZrNWREwCiIjXI+LNmvWwMuX890LSUGBz4PbadKsqOh1bRDwREU+m4+fJ/hjp9Asl6qCc94P8eK8HPtLZioiDWy+RpuF7AA/VuSvVsBXwXO58AQ0etJPNI2JROv4b2RviKiLiAeBuYFF63RYR82rXxYp0Oi5gEPB3STdIeljS+ekv7kbW6bgkrQFcAIypZceqoJz/Zv8iaS+yP7j+2tMdq0A57wf/yhMRy4BXgc06qtRfv9WDJN0BvLeNS9+JiJu7UE9fYAJwRkT8o1r9q1S1xtWIOhpb/iQiQtK7Pkcj6QPAjsDWKWmSpP0j4t6qd7YLujsusveK/cn+wHoWuBYYCfymuj3tmiqM66vAHyNiQaPdGq3C2Frq2QK4AvhCRKyobi8bl4NbD4qIQ7tbh6S1yQLbVRFxQ/d71X1VGNdCYJvc+dYpre46GpukFyRtERGL0htGW/ecjgIejIjXU5k/AXsDdQ1uVRjXAmBmRDyVytxEdh+4rsGtCuPaG9hf0leBvsA6kl6PiLo/5FSFsSFpI+AWsj88H+yhrnZXOe8HLXkWSFoL6Ae83FGlXpZsYGlN+TfAvIj4ab37U0XTgO0lvU/SOsBxwMQ696kcE4EvpOMvAG3NUp8FDpS0VvrD5ECye6WNrJxxTQM2ltRyz+YQYG4N+tYdnY4rIo6PiIER0US2NPnbRghsZeh0bOn/rRvJxnR9DfvWVeW8H+THewxwV3T2DSQR4VcdXmR/4S8AlpI9cHBbSt+SbJkEYD+yBxRmAzPT62P17nt3x5XOPwY8QXYP4Dv17neZY9uM7Mm0J4E7gE1Tegm4JB2vCfyKLKDNBX5a735XY1zpfFj6tzgHGA+sU+++V2NcufwjgbH17ne1xgZ8Dngn994xExhS7763M553vR8A3wc+lY77ANcBfwGmAu/vrE5//ZaZmRWOlyXNzKxwHNzMzKxwHNzMzKxwHNzMzKxwHNzMzKxwHNzMzKxwHNzMzKxw/j/6HZ7QCtPmHwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.851474940776825\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.9393419623374939\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbcAAAEICAYAAADcC7pLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAnEElEQVR4nO3debxd873/8ddbhCDEkFBTnE6JWcimNccQV3VA0ZiqQZurWtqkqFv9tVq9qtc11KVDKKEUrTGt1jy3SE4ik0S05pASoiqGlOTz+2N9T60cZ9hnn3323mfl/Xw89uOs9V3f8Yj9Od/vWnt/FRGYmZkVyQr17oCZmVm1ObiZmVnhOLiZmVnhOLiZmVnhOLiZmVnhOLiZmVnhOLiZFZik70i6pN79MKs1+XNuZm2T9AywHrAklzwkIl7sZp1fjog7u9e73kfS6cDHIuLIevfFis8zN7OOfTYi+udeFQe2apC0Yj3br1Rv7bf1Xg5uZl0kaYCkX0maL+kFST+S1Cdd+6ikuyW9KukVSVdJWjNd+zUwGPi9pEWSTpE0QtK8VvU/I2nvdHy6pOskXSnpn8Dojtpvo6+nS7oyHTdJCklHS3pe0muSjpO0vaQZkv4h6cJc2dGS/izpQkmvS3pc0l656xtImihpoaS/SfpKq3bz/T4O+A4wKo19esp3tKQ5kt6Q9JSk/8zVMULSPEnfkvRyGu/RueurSDpH0rOpfw9KWiVd+6Skv6QxTZc0ooL/1NaLObiZdd0E4D3gY8C2wD7Al9M1AT8GNgA2AzYGTgeIiC8Cz/H+bPB/ymxvf+A6YE3gqk7aL8cngI8Do4DzgdOAvYEtgC9I2r1V3ieBgcD3gRskrZ2uXQPMS2M9GDhT0p7t9PtXwJnAtWns26Q8LwOfAdYAjgbOk7Rdro4PAQOADYFjgYskrZWu/S8wHNgJWBs4BVgqaUPgFuBHKf0k4HpJg7rwO7JezsHNrGM3pb/+/yHpJknrAfsB34yINyPiZeA84FCAiPhbRNwREYsjYgFwLrB7+9WX5aGIuCkilpIFgXbbL9MZEfFORNwOvAlcHREvR8QLwANkAbPFy8D5EfFuRFwLzAU+LWljYGfg26muacAlwFFt9Tsi3m6rIxFxS0Q8GZn7gNuBXXNZ3gV+mNr/I7AIGCppBeAY4BsR8UJELImIv0TEYuBI4I8R8cfU9h1Ac/q92XLC6+BmHTsg//CHpB2AvsB8SS3JKwDPp+vrAT8le4NePV17rZt9eD53vElH7Zfppdzx222c98+dvxDLPnX2LNlMbQNgYUS80epaqZ1+t0nSp8hmhEPIxrEqMDOX5dWIeC93/lbq30CgH9mssrVNgEMkfTaX1he4p7P+WHE4uJl1zfPAYmBgqzfdFmcCAWwVEQslHQBcmLve+vHkN8ne0AFI985aL5/ly3TWfrVtKEm5ADcYmAi8CKwtafVcgBsMvJAr23qsy5xLWhm4nmy2d3NEvCvpJrKl3c68ArwDfBSY3ura88CvI+IrHyhlyw0vS5p1QUTMJ1s6O0fSGpJWSA+RtCw9rk62dPZ6uvdzcqsqXgI+kjt/Augn6dOS+gLfBVbuRvvVti5woqS+kg4hu4/4x4h4HvgL8GNJ/SRtTXZP7MoO6noJaEpLigArkY11AfBemsXtU06n0hLtpcC56cGWPpJ2TAHzSuCzkv4jpfdLD6ds1PXhW2/l4GbWdUeRvTHPJltyvA5YP137AbAd8DrZQw03tCr7Y+C76R7eSRHxOnA82f2qF8hmcvPoWEftV9sjZA+fvAL8N3BwRLyarh0GNJHN4m4Evt/J5/d+l36+KmlqmvGdCPyWbByHk80Ky3US2RLmZGAh8BNghRR49yd7OnMB2UzuZPx+t1zxh7jNrE2SRpN94HyXevfFrKv8l4yZmRWOg5uZmRWOlyXNzKxwPHMzM7PC8efcamDgwIHR1NRU726YmfUqU6ZMeSUiKvraNAe3GmhqaqK5ubne3TAz61UkPVtpWS9LmplZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4fhD3L3QeXc8Ue8umJk1NM/czMyscBzczMyscBzczMyscBzczMyscKoS3CQtkTRN0nRJUyXtVIU6h0naL3c+WtKC1M40SVd0Uv5eSaV0/IykgZX0VdKako7PnY+Q9Ifujc7MzHpStWZub0fEsIjYBvgv4MdVqHMYsF+rtGtTO8Mi4qgK6+1qX9cEju8kj5mZNZCeWJZcA3gNQNL6ku5PM6VZknZN6YsknS3pMUl3StohzbSekvQ5SSsBPwRGpbKj2mqo9SxK0oWSRlfY1/6S7kqzuZmS9k95zgI+mvpxdkrrL+k6SY9LukqSutCmmZn1sGp9zm0VSdOAfsD6wJ4p/XDgtoj4b0l9gFVT+mrA3RFxsqQbgR8BI4HNgcsjYqKk7wGliPg6ZMuSZMFul1THT4Gnq9jXd4ADI+KfaQnzYUkTgVOBLSNiWOrHCGBbYAvgReDPwM7Ag/lGJI0BxgAMHjy4gm6amVmlqhXc3s69+e8IXCFpS2AycKmkvsBNETEt5f8XcGs6ngksjoh3Jc0Emjpo59qWYJfaGlHFvgo4U9JuwFJgQ2C9duqYFBHzUh3TUp+XCW4RMR4YD1AqlaKCfpqZWYWqviwZEQ8BA4FBEXE/sBvwAjBBUst9sncjouUNfymwOJVdStcC7nssO4Z+lfYVOCL9HJ6C30sd1Lc4d7wEf9OLmVlDqXpwk7Qp0Ad4VdImwEsRcTFwCbBdF6p6A1i9kzzPAptLWlnSmsBelfYVGAC8nGaQewCbdKEfZmbWQKp9zw2y5b0vRcSStGx4sqR3gUVAV55wvAc4NdXb5hONEfG8pN8Cs8juvz3ajb5eBfw+LY02A4+nNl6V9GdJs4A/Abd0YQxmZlYHen910HpKqVSK5ubmqtXnL042s+XBuH2GTomIUiVl/Q0lZmZWOA5uZmZWOH7KrxcaO3JIvbtgZtbjxnWjrGduZmZWOA5uZmZWOA5uZmZWOL7n1gv5owBmZh3zzM3MzArHwc3MzArHwc3MzArHwc3MzAqnYYKbpCVpt+uW16md5P9Ohe2sJOl8SX9Lrz9Iqng3UUmnSzqp0vJmZlZ9jfS05L83ES3Td4Azu9JA2g38TLItbIam3QCOBm6WNDztJ2dmZr1cw8zc2iJpgKS5koam86slfUXSWaSta9JWNUg6UtKklPbLFMiQtEjSOZKmAzsDRwNjI2IJQERcRrYdz96SmtLWNi3tnyTp9HT8FUmTJU2XdL2kVWv4qzAzsy5opODWEqxaXqMi4nXg62S7eB8KrBURF0fEqaSZXkQcIWkzYBSwc5r9LSHbWRtgNeCRiNgG+AfwXET8s1XbzcDmnfTvhojYPtUzBzi2o8ySxkhqltS8YMGCsn8JZmbWfQ2/LBkRd0g6BLgI2KadsnsBw4HJkgBWAV5O15YA11ehf1tK+hGwJtAfuK2jzBExHhgP2X5uVWjfzMzK1EjBrU2SVgA2A94C1gLmtZUNuDwi/quNa++0LEECTwKDJa0eEW/k8gwnC4Dvsexstl/ueAJwQERMlzQaGNH10ZiZWS000rJke8aSLQMeDlwmqW9Kfzd3fBdwsKR1ASStLWmT1hVFxJvA5cC5uXtyRwHvAH8GXgLWlbSOpJWBz+SKrw7MT20egZmZNaxGmrmtImla7vxW4DLgy8AOEfGGpPuB7wLfJ1vymyFparrv9l3g9jTTexf4GvBsG+38F3A2MFfSKsACYMeICLKA+UNgEvAC8Hiu3P8DHkn5HyELdmZm1oCUvacvnyR9CPgT8PN0j6xHlEqlaG5urlp9/uJkM1sejNtn6JSIKFVStpFmbjUXEX8Htq13P8zMrLp6wz03MzOzLnFwMzOzwlmulyV7q7Ejh9S7C2ZmPW5cN8p65mZmZoXj4GZmZoXj4GZmZoXje269kD/nZmbWMc/czMyscBzczMyscBzczMyscCoKbpJC0jm583/vWN1Bmc9JOrWTPCMk/aGda89IGlhJf1P5CZIOrrR8res1M7PKVTpzWwx8vivBJiImRsRZFbbXLZL84IyZ2XKk0uD2HtmWM2NbX5A0SNL1kian184pfbSkC9PxRyU9LGmmpB9JWpSror+k6yQ9Lukqpa21k1NSmUmSPpbqapJ0t6QZku6SNDilT5D0C0mPAP+Tyu8m6S+SnmqZbSlztqRZqe5RZaRfKGmupDuBdSv8HZqZWQ/pzj23i4AjJA1olf5T4LyI2B44CLikjbI/BX4aEVvxwZ21twW+CWwOfATYOXft9VTmQuD8lPZ/ZLtwbw1cBVyQy78RsFNEtHyLy/rALmSbkLbMIj8PDAO2AfYGzpa0fgfpBwJDU/+OAnZqY3xmZlZHFQe3iPgncAVwYqtLewMXpo1HJwJrSOrfKs+OwO/S8W9aXZsUEfMiYikwDWjKXbs693PHXF0tdfyaLHi1+F1ELMmd3xQRSyNiNrBeStsFuDoilkTES8B9wPYdpO+WS38RuJs2SBojqVlS84IFC9rKYmZmPaS796LOB6aS7ZjdYgXgkxHxTj7jsquLHVqcO17Csn2Mdo7b82YHdZfdoUqkzU/HQ7ZZaU+2ZWZmy+rWRwEiYiHwW+DYXPLtwAktJ5KGtVH0YbIlS4BDu9DkqNzPh9LxX3J1HAE80IX6SPlHSeojaRDZzGxSB+n359LXB/boYntmZtbDqvEU4TnA13PnJwIXSZqR6r8fOK5VmW8CV0o6DbgVeL3MttZK9S4GDktpJwCXSToZWAAc3cX+30i2tDmdbDZ4SkT8XVJH6XsCs4HneD/ImplZg1BE7VfMJK0KvB0RIelQ4LCI2L/mHamRUqkUzc3NVavP3y1pZsuDcfsMnRIRpUrK1uvzX8PJHjoR8A/gmDr1w8zMCqguwS0iHiB7xN7MzKzq/N2SZmZWOP5aql5o7Mgh9e6CmVmPG9d5lnZ55mZmZoXj4GZmZoXj4GZmZoXje269kD/nZmbWMc/czMyscBzczMyscBzczMyscBzczMyscHokuElaR9K09Pq7pBdy5yu1yvvN9EXKndV5r6RSOn5G0sxU30xJ3f7SZUlNkg7Pna8q6apU/yxJD7ZsuippSW480yQ1dbd9MzOrnh55WjIiXgWGAUg6HVgUEf/bTvZvAlcCb3WxmT0i4hVJQ8n2kLu5os6+rwk4nPd39f4G8FJEbAWQ2nk3XXs7IoZ1sz0zM+shNVuWlLSXpEfTTOhSSStLOhHYALhH0j0p388lNUt6TNIPyqh6DeC1VHY1SbdImp5mW6NS+jOSfpxmWc2StpN0m6QnJbXsNXcWsGvKMxZYH3ihpZGImBsRizEzs4ZXq8+59QMmAHtFxBOSrgC+GhHnSxpHmoWlvKdFxEJJfYC7JG0dETPaqPOetGXOR4AvpLR9gRcj4tMAkgbk8j8XEcMknZf6snPq1yzgF8CpwEkR8ZlUdhhwu6SDgbuAyyPir6muVSRNS8dPR8SBrTsnaQwwBmDw4MFl/6LMzKz7ajVz60MWBFo+fXw5sFs7eb8gaSrwKLAFsHk7+faIiC2Brcj2husPzARGSvqJpF0jIr/D98T0cybwSES8ERELgMWS1mxdeURMIwucZwNrA5MlbZYuvx0Rw9LrA4EtlR8fEaWIKA0aNKidIZiZWU9oqKclJX0YOIlshrc1cAvZ7KpdEfEk8BKweQqe25EFsB9J+l4ua8uS4tLccct5mzPYiFgUETdExPFk9wX36/qozMys1moV3JYATZI+ls6/CNyXjt8AVk/HawBvAq9LWg/4VGcVS1oX+DDwrKQNgLci4kqyGdd2Xehjvh9I2lnSWul4JbIZ5LNdqM/MzOqkVvfc3gGOBn4naUVgMtl9LoDxwK2SXoyIPSQ9CjwOPA/8uYM675G0BOgLnBoRL0n6D+BsSUvJnmz8ahf6OANYImk62T25V4Gfp/t6K5DNIq/vQn1mZlYnioh696HwSqVSNDc3V60+f3GymS0Pxu0zdEpElCop21D33MzMzKrBwc3MzArH+7n1QmNHDql3F8zMety4bpT1zM3MzArHwc3MzArHwc3MzArH99x6IX8UwMysY565mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4XQY3JR5UNKncmmHSLq1O41KWpJ2vJ4l6fdt7afWUySNlnRhq7Rpkq7poMwISX9o59ozkgZWu59mZla5DoNbZN+qfBxwrqR+aUPQM4GvVdJY2hEA3t/sc0tgYaX1VUPagLQPsKuk1erVDzMzq55OlyUjYhbwe+DbwPfINu08TdIkSY9K2h9AUpOkByRNTa+dUvqIlD4RmN1GEw8BG6a8H5V0q6QpqcymKX2CpJ9LeljSU6nOSyXNkTShpSJJh0mamWaEP8mlHy3pCUmTgJ1btX8Y8GvgdmD/XJl9JT2edgX/fC59HUm3S3pM0iWAOvsdmplZbZV7z+0HwOFkm4f2A+6OiB2APcj2T1sNeBkYGRHbAaOAC3LltwO+ERHLfCmipD7AXsDElDQeOCEihpPtyP2zXPa1gB2BsSn/ecAWwFaShqWNSn8C7AkMA7aXdICk9VP/dwZ2Idt0NG8UcA1wNVmgQ1I/4GLgs8Bw4EO5/N8HHoyILYAbgcFt/cIkjZHULKl5wYIFbWUxM7MeUtaHuCPiTUnXAouALwCflXRSutyP7A3+ReBCScPIdt7OB7JJEfF07nwVSdPIZmxzgDvSkudOZBuatuRbOVfm9xERkmYCL0XETABJjwFNwCbAvRGxIKVfBeyWyubTr23pm6QS8EpEPCfpBeBSSWun8TwdEX9N+a4ExqS6diPN5CLiFkmvtfM7G08WrCmVSt40z8yshrryDSVL00vAQRExN39R0unAS8A2ZDPCd3KX32xV19sRMUzSqsBtZPfcJgD/iIhh7bS/ONePxbn0pWkc73ZhLC0OAzaV9Ew6XwM4iGyncDMz66Uq+SjAbcAJStMrSdum9AHA/IhYCnyR7CGNDkXEW8CJwLeAt4CnJR2S6pWkbbrQr0nA7pIGpuXOw4D7gEdS+jqS+gIt9a9ANgvdKiKaIqKJ7J7bYcDjQJOkj6a6D8u1cz/ZEi3pKdK1utBHMzOrgUqC2xlAX2BGWhI8I6X/DPiSpOnApnxwttamiHgUmEEWQI4Ajk11PEbuAY8y6pkPnArcA0wHpkTEzSn9dLIHV/5MtgwKsCvwQkS8mKvmfrJ7cmuRLUPekh4oeTmX5wfAbmnsnweeK7ePZmZWG8qe9reeVCqVorm5uWr1+YuTzWx5MG6foVMiolRJWX9DiZmZFY6Dm5mZFY73c+uFxo4c0nkmM7Neblw3ynrmZmZmhePgZmZmhePgZmZmheN7br2QPwpgZtYxz9zMzKxwHNzMzKxwHNzMzKxwHNzMzKxwejy4SVoiaVru1STpL1Ws/xlJA6tVX6pzl7TT+OOS5ko6vpr1m5lZz6rF05Jvt7FH206tM0laMSLeq0F/OiTpQ8BvgAMiYmoKnLdJmh8RN9a5e2ZmVoa6LEtKWpR+jpD0gKSJwGxJfSSdLWmypBmS/jOX735Jt6SZ1C/Sfmyt671J0hRJj0kak0vfV9JUSdMl3ZXSVpN0aZqhPSqpZXudrwETImIqQES8ApwCnJzKTZB0cOuxmJlZ46jFzG0VSdPS8dMRcWCr69sBW0bE0ykgvR4R20taGfizpNtTvh3I9lp7FriVbC+161rVdUxELJS0CjBZ0vVkAfxiYLfUxtop72nA3RFxjKQ1gUmS7gS2AC5vVW9zartsaSxjAAYPHtyVomZm1k31WpbMmxQRT6fjfYCtczOjAcDHgX+lfE8BSLoa2IUPBrcTJbUEz41T2UHA/S1tRMTCXFufk3RSOu8HVC0KRcR4YDxk+7lVq14zM+tcI3xDSX7HbgEnRMRt+QySRgCtA0S0kWdvYMeIeEvSvWQBqz0CDoqIua3qmQ0MB27OJQ8nm70BvEdazk1Loyt10IaZmdVBo30U4Dbgq5L6AkgaImm1dG0HSR9OAWUU8GCrsgOA11Jg2xT4ZEp/GNhN0odTnS3LkrcBJ0hSSt82pV8EjJY0LKWvA/w3cEa6/gxZsAP4HNC326M2M7OqarTgdgkwG5gqaRbwS96fXU4GLgTmAE8DrZ9cvBVYUdIc4CyyoEZELCC793WDpOnAtSn/GWSBaYakx9I5ETEfOBIYL2ku8CJwQUTcl8pdDOye6tqRZWeeZmbWABTR+LeD0pLjSRHxmTq0fTzwVbIHUl6rpI5SqRTNzc2dZyyTvzjZzJYH4/YZOiUiSpWUbbSZW8OJiJ9FxFaVBjYzM6u9RnigpFMRcS9wb527YWZmvUSvCG62rLEjh9S7C2ZmPW5cN8p6WdLMzArHwc3MzArHwc3MzArH99x6IX8UwMysY565mZlZ4Ti4mZlZ4Ti4mZlZ4Ti4mZlZ4dQkuEnaSNLNkv4q6UlJP5XUo1vF5Hb7bkpfwtySvkvaffvxtKv38dVox8zMGkePB7e0pcwNwE0R8XFgCNCfbBuZ7tTb5Sc9JX0I+A1wXERsCuwMHJvb4NTMzAqgFjO3PYF3IuIygIhYAowFjkkzqC1aMkq6V1JJ0mqSLk3XH5W0f7o+WtJESXcDd0nqL+kuSVMlzWzJ14GvARMiYmrqyyvAKcDJqf4JuV3A87O/rrZjZmZ1VIvPuW0BTMknRMQ/JT0H3AJ8Afi+pPWB9SOiWdKZwN0RcYykNYFJku5MxbcDto6IhWn2dmCqbyDwsKSJ0f4+PlsAl7dKawY272QM73SxHSSNIdtHjsGDB3dSvZmZVVO9Hyi5F2iZKX0BuC4d7wOcKmlaytMPaIkQd0TEwnQs4ExJM4A7gQ2B9Xqgn11uJyLGR0QpIkqDBg3qgS6ZmVl7ajFzm837AQwASWuQBavJwKuStgZGAce1ZAEOioi5rcp9gmV3vj4CGAQMj4h3JT1DFgg76stw4OZc2nCy2RvAe6SAL2kFoOWhl662Y2ZmdVSLmdtdwKqSjgKQ1Ac4h+ze11vAtWT3vQZExIxU5jbghPQwCpK2bafuAcDLKeDsAWzSSV8uAkZLGpbqXYfswZYz0vVnyIIdwOeAvhW2Y2ZmddTjwS3dlzoQOETSX4EnyO5hfSdluQ44FPhtrtgZZIFlhqTHeD/4tHYVUJI0EzgKeLyTvswHjgTGS5oLvAhcEBH3pSwXA7tLmg7syPuzxC61Y2Zm9aUOnokovPQZt68Cu0XEaz3VTqlUiubm5s4zlslfnGxmy4Nx+wydEhGlSsrW+4GSuoqIn0XEVj0Z2MzMrPaW6+BmZmbF5P3ceqGxI4fUuwtmZj1uXDfKeuZmZmaF4+BmZmaF4+BmZmaF43tuvYg/AmBmVh7P3MzMrHAc3MzMrHAc3MzMrHAc3MzMrHAaOrhJ2ljSPZJmS3pM0je6WP5eSaV0/EzaRXtaeu0kqUnSrHbKriDpAkmzUrnJkj7cXl3dH62ZmVVLoz8t+R7wrYiYKml1YIqkOyJidoX17RERr7ScSGpqK1Pa4fsQYAOyXb+XStqIZfeSW6YuMzNrHA0d3NIWNfPT8RuS5gAbSvoZ8AiwB7AmcGxEPCBpFeAyYBuybWlWKbctSaOBzwP9gT5kG5rOj4ilqf15VRqWmZn1sIYObnlplrUtWVADWDEidpC0H/B9YG+y7WveiojN0u7eU1tVc4+kJcDiiPhEG81sRzZTW5hmag9K2pVsw9UrI+LRLtRlZmZ10iuCm6T+wPXANyPin2mD7hvS5SlAUzreDbgAICJmSJrRqqrOlhLviIiFqfw8SUOBPdPrLkmHRMRd5dQlaQwwBmDw4MHlDdTMzKqioR8oAZDUlyywXRURN+QuLU4/l1C9IJ2/p0ZELI6IP0XEycCZwAHlVhQR4yOiFBGlQYMGVal7ZmZWjoYObsqmaL8C5kTEuWUUuR84PJXdEti6G21vJ2mDdLxCquvZSuszM7PaafRlyZ2BLwIzJU1Lad/pIP/PgcvSgydzyJYsK7UucLGkldP5JODCbtRnZmY10tDBLSIeBNTGpT/m8rxCuucWEW8Dh7ZTV1Mbac8AW6bjCcCE3LVbgVvLrcvMzBpHQy9LmpmZVcLBzczMCsfBzczMCqeh77nZssaOHFLvLpiZ1cy4bpT1zM3MzArHwc3MzArHwc3MzArH99x6kfPueKLeXTAz6xU8czMzs8JxcDMzs8JxcDMzs8JxcDMzs8LpVcFN0qWSXpY0q5N8IyTtlDs/XdILkqal11kp/V5JpXbq+IykRyVNlzRb0n92VJeZmTWO3va05ASybWeu6CTfCGAR8Jdc2nkR8b/lNJK2uRkP7JB25F6Z93f77lJdZmZWe71q5hYR9wML82mSTkwzqxmSrpHUBBwHjE0zq13LqVvSIknnSJoOfIIs8L+a2l0cEXOrOhgzM+sxvSq4teNUYNuI2Bo4Lu3R9guy2dWwiHgg5RubW0r8jzbqWQ14JCK2SUF0IvCspKslHZF246bMupA0RlKzpOYFCxZUbbBmZta5IgS3GcBVko4E3usgX0uwGxYRt7VxfQlwfctJRHwZ2ItsB+6TgEu7UBcRMT4iShFRGjRoUFfHZGZm3VCE4PZp4CJgO2CypErvI74TEUvyCRExMyLOA0YCB3Wvm2ZmViu9OrilpcKNI+Ie4NvAAKA/8Aawejfq7S9pRC5pGPBsxR01M7Oa6lVPS0q6muxJyIGS5gFnAF+UNAAQcEFE/EPS74HrJO0PnFBJU8Apkn4JvA28CYyuwhDMzKwGelVwi4jD2kj+ZRv5ngC2ziU90DpPyjcid9w/d/wGsF87ZU4vr7dmZlYvvXpZ0szMrC0ObmZmVji9allyeTd25JB6d8HMrGbGdaOsZ25mZlY4Dm5mZlY4Dm5mZlY4vufWC5x3xxP17oKZWa/imZuZmRWOg5uZmRWOg5uZmRWOg5uZmRVOQwU3SetJ+o2kpyRNkfSQpAPbyNckaVYb6T+UtHcZ7QyTFJL2rVbfzcyscTRMcJMk4Cbg/oj4SEQMBw4FNmqVr90nPCPiexFxZxnNHQY8mH622ZdWO2+bmVkv0khv4HsC/4qIX7QkRMSzEfF/kkZLmijpbuCu9iqQNEHSwZL2lfS7XPoISX9IxwIOIdvCZqSkfim9SdJcSVcAs4CNJZ0sabKkGZJ+kKvvpjSzfEzSmOr+GszMrLsaKbhtAUzt4Pp2wMERsXsZdd0JfELSaul8FHBNOt4JeDoingTuJdvJu8XHgZ9FxBbA0HS+A9lmpcMl7ZbyHZNmliXgREnrtO6ApDGSmiU1L1iwoIwum5lZtTRScFuGpIskTZc0OSXdERELyykbEe8BtwKfTcuYnwZuTpcP4/1Adw3LLk0+GxEPp+N90utRsqC7KVmwgyygTQceBjbOpef7MD4iShFRGjRoUDndNjOzKmmkbyh5DDio5SQiviZpINCckt7sYn3XAF8HFgLNEfGGpD6pjf0lnUa24/Y6klZvow0BP46IZTZDlTQC2BvYMSLeknQv0K+LfTMzsx7USDO3u4F+kr6aS1u1G/XdR7aU+RXen6ntBcyIiI0joikiNgGuBz7wRCZwG3CMpP4AkjaUtC4wAHgtBbZNgU92o49mZtYDGia4RUQABwC7S3pa0iTgcuDb7RQZKmle7nVIq/qWAH8APpV+QrYEeWOreq6njacmI+J24DfAQ5JmAtcBq5Mtd64oaQ5wFtnSpJmZNRBlMcV6UqlUiubm5s4ztsNfnGxmy6Nx+wydEhGlSso2zMzNzMysWhzczMyscBrpaUlrx9iRQ+rdBTOzmhvXjbKeuZmZWeE4uJmZWeE4uJmZWeE4uJmZWeE4uJmZWeE4uJmZWeE4uJmZWeE4uJmZWeE4uJmZWeH4i5NrQNIC4Nlc0kDglTp1p6d5bL1TkccGxR5fkcc2NCJW7zzbB/nrt2ogIpbZiltSc6XfdN3oPLbeqchjg2KPr+hjq7SslyXNzKxwHNzMzKxwHNzqY3y9O9CDPLbeqchjg2KPz2Nrgx8oMTOzwvHMzczMCsfBzczMCsfBrQYkHSLpMUlLJbX5yK6kjSXdI2l2yvuNWvezEuWMLeXbV9JcSX+TdGot+1gpSWtLukPSX9PPtdrJ9z/pdzBH0gWSVOu+dlUXxjZY0u1pbLMlNdW4qxUpd3wp7xqS5km6sJZ9rFQ5Y5M0TNJD6d/lDEmj6tHXcnX2/iBpZUnXpuuPlPPv0MGtNmYBnwfu7yDPe8C3ImJz4JPA1yRtXovOdVOnY5PUB7gI+BSwOXBYLxnbqcBdEfFx4K50vgxJOwE7A1sDWwLbA7vXspMV6nRsyRXA2RGxGbAD8HKN+tdd5Y4P4Aw6/n+z0ZQztreAoyJiC2Bf4HxJa9aui+Ur8/3hWOC1iPgYcB7wk87qdXCrgYiYExFzO8kzPyKmpuM3gDnAhrXoX3eUMzayN8W/RcRTEfEv4Bpg/57vXbftD1yeji8HDmgjTwD9gJWAlYG+wEu16Fw3dTq29AazYkTcARARiyLirZr1sHvK+W+HpOHAesDttelWVXQ6toh4IiL+mo5fJPujZFDrfA2inPeH/JivA/bqbIXEwa0BpSn3tsAjde5KtWwIPJ87n0cvCNzAehExPx3/nexNcBkR8RBwDzA/vW6LiDm162LFOh0bMAT4h6QbJD0q6ez0V3Zv0On4JK0AnAOcVMuOVUE5/+3+TdIOZH98PdnTHatQOe8P/84TEe8BrwPrdFSpv36rSiTdCXyojUunRcTNXainP3A98M2I+Ge1+tcd1RpbI+pobPmTiAhJH/jcjKSPAZsBG6WkOyTtGhEPVL2zXdTdsZG9P+xK9ofWc8C1wGjgV9XtaWWqML7jgT9GxLxGu01ahbG11LM+8GvgSxGxtLq9bGwOblUSEXt3tw5JfckC21URcUP3e1UdVRjbC8DGufONUlrddTQ2SS9JWj8i5qc3ibbuNx0IPBwRi1KZPwE7AnUPblUY2zxgWkQ8lcrcRHY/uCGCWxXGtyOwq6Tjgf7ASpIWRUTdH3iqwtiQtAZwC9kfoQ/3UFeroZz3h5Y88yStCAwAXu2oUi9LNoi0fvwrYE5EnFvv/lTZZODjkj4saSXgUGBinftUjonAl9Lxl4C2ZqnPAbtLWjH9cbI72f3SRlfO2CYDa0pquVezJzC7Bn2rhk7HFxFHRMTgiGgiW5q8ohECWxk6HVv6/+xGsjFdV8O+VaKc94f8mA8G7o7OvoEkIvzq4RfZX/fzgMVkDxvcltI3IFsWAdiF7OGEGcC09Nqv3n2vxtjS+X7AE2Tr/qfVu99ljm0dsqfR/grcCayd0kvAJem4D/BLsoA2Gzi33v2u1tjS+cj0b3ImMAFYqd59r+b4cvlHAxfWu9/VGhtwJPBu7r1kGjCs3n3vYEwfeH8Afgh8Lh33A34H/A2YBHykszr99VtmZlY4XpY0M7PCcXAzM7PCcXAzM7PCcXAzM7PCcXAzM7PCcXAzM7PCcXAzM7PC+f/uqTOAGq84IwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.825083315372467\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.8057377338409424\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.3722544312477112\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.1210969686508179\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-1.0212703943252563\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.7191378474235535\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.03590421751141548\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.07192059606313705\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.3259488046169281\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6218269467353821\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.24499619007110596\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.36947116255760193\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-0.6966938376426697\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.28163987398147583\n"
     ]
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 45;\n",
       "                var nbb_unformatted_code = \"for i in range(20):\\n    explain_plot(explanations[i], all_cols)\\n    print(probs[i].item())\";\n",
       "                var nbb_formatted_code = \"for i in range(20):\\n    explain_plot(explanations[i], all_cols)\\n    print(probs[i].item())\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range(20):\n",
    "    explain_plot(explanations[i], all_cols)\n",
    "    print(probs[i].item())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:38.516694Z",
     "start_time": "2020-09-19T17:12:38.494176Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 46;\n",
       "                var nbb_unformatted_code = \"from sklearn.metrics import confusion_matrix\";\n",
       "                var nbb_formatted_code = \"from sklearn.metrics import confusion_matrix\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:38.808574Z",
     "start_time": "2020-09-19T17:12:38.517685Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 47;\n",
       "                var nbb_unformatted_code = \"probs_test, explanations_train = predict(model, input_test)\";\n",
       "                var nbb_formatted_code = \"probs_test, explanations_train = predict(model, input_test)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "probs_test, explanations_train = predict(model, input_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:38.958480Z",
     "start_time": "2020-09-19T17:12:38.809711Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 48;\n",
       "                var nbb_unformatted_code = \"# plt.barh(y_pos, performance, align=\\\"center\\\", alpha=0.5)\\n# plt.yticks(y_pos, columns[selection])\\nplt.ylabel(\\\"Predicted value\\\")\\nplt.xlabel(\\\"Truth value\\\")\\n\\nplt.title(\\\"Scatter plot of prediction vs truth\\\")\\nplt.scatter(\\n    train[target].values[test_indices], target_encoder.inverse_transform(probs_test),\\n)\\nplt.plot(\\n    [\\n        np.min(train[target].values[test_indices]),\\n        np.max(train[target].values[test_indices]),\\n    ],\\n    [\\n        np.min(train[target].values[test_indices]),\\n        np.max(train[target].values[test_indices]),\\n    ],\\n)\\nplt.grid(True)\\nplt.show()\";\n",
       "                var nbb_formatted_code = \"# plt.barh(y_pos, performance, align=\\\"center\\\", alpha=0.5)\\n# plt.yticks(y_pos, columns[selection])\\nplt.ylabel(\\\"Predicted value\\\")\\nplt.xlabel(\\\"Truth value\\\")\\n\\nplt.title(\\\"Scatter plot of prediction vs truth\\\")\\nplt.scatter(\\n    train[target].values[test_indices],\\n    target_encoder.inverse_transform(probs_test),\\n)\\nplt.plot(\\n    [\\n        np.min(train[target].values[test_indices]),\\n        np.max(train[target].values[test_indices]),\\n    ],\\n    [\\n        np.min(train[target].values[test_indices]),\\n        np.max(train[target].values[test_indices]),\\n    ],\\n)\\nplt.grid(True)\\nplt.show()\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plt.barh(y_pos, performance, align=\"center\", alpha=0.5)\n",
    "# plt.yticks(y_pos, columns[selection])\n",
    "plt.ylabel(\"Predicted value\")\n",
    "plt.xlabel(\"Truth value\")\n",
    "\n",
    "plt.title(\"Scatter plot of prediction vs truth\")\n",
    "plt.scatter(\n",
    "    train[target].values[test_indices],\n",
    "    target_encoder.inverse_transform(probs_test),\n",
    ")\n",
    "plt.plot(\n",
    "    [\n",
    "        np.min(train[target].values[test_indices]),\n",
    "        np.max(train[target].values[test_indices]),\n",
    "    ],\n",
    "    [\n",
    "        np.min(train[target].values[test_indices]),\n",
    "        np.max(train[target].values[test_indices]),\n",
    "    ],\n",
    ")\n",
    "plt.grid(True)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.096220Z",
     "start_time": "2020-09-19T17:12:38.959684Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 49;\n",
       "                var nbb_unformatted_code = \"plt.hist(\\n    train[target].values[test_indices].reshape(-1)\\n    - target_encoder.inverse_transform(probs_test).reshape(-1),\\n)\\nplt.xlabel(\\\"Error value\\\")\\nplt.ylabel(\\\"Counts\\\")\\nplt.show()\";\n",
       "                var nbb_formatted_code = \"plt.hist(\\n    train[target].values[test_indices].reshape(-1)\\n    - target_encoder.inverse_transform(probs_test).reshape(-1),\\n)\\nplt.xlabel(\\\"Error value\\\")\\nplt.ylabel(\\\"Counts\\\")\\nplt.show()\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(\n",
    "    train[target].values[test_indices].reshape(-1)\n",
    "    - target_encoder.inverse_transform(probs_test).reshape(-1),\n",
    ")\n",
    "plt.xlabel(\"Error value\")\n",
    "plt.ylabel(\"Counts\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.222366Z",
     "start_time": "2020-09-19T17:12:39.097164Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 50;\n",
       "                var nbb_unformatted_code = \"plt.hist(\\n    np.abs(\\n        train[target].values[test_indices].reshape(-1)\\n        - target_encoder.inverse_transform(probs_test).reshape(-1),\\n    )\\n)\\nplt.xlabel(\\\"Error value (absolute)\\\")\\nplt.ylabel(\\\"Counts\\\")\\nplt.show()\";\n",
       "                var nbb_formatted_code = \"plt.hist(\\n    np.abs(\\n        train[target].values[test_indices].reshape(-1)\\n        - target_encoder.inverse_transform(probs_test).reshape(-1),\\n    )\\n)\\nplt.xlabel(\\\"Error value (absolute)\\\")\\nplt.ylabel(\\\"Counts\\\")\\nplt.show()\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(\n",
    "    np.abs(\n",
    "        train[target].values[test_indices].reshape(-1)\n",
    "        - target_encoder.inverse_transform(probs_test).reshape(-1),\n",
    "    )\n",
    ")\n",
    "plt.xlabel(\"Error value (absolute)\")\n",
    "plt.ylabel(\"Counts\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.250809Z",
     "start_time": "2020-09-19T17:12:39.223544Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 51;\n",
       "                var nbb_unformatted_code = \"out = Path(os.getcwd() + \\\"/data/\\\" + dataset_name + \\\"-test.csv\\\")\";\n",
       "                var nbb_formatted_code = \"out = Path(os.getcwd() + \\\"/data/\\\" + dataset_name + \\\"-test.csv\\\")\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "out = Path(os.getcwd() + \"/data/\" + dataset_name + \"-test.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.286863Z",
     "start_time": "2020-09-19T17:12:39.251763Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 52;\n",
       "                var nbb_unformatted_code = \"kaggle_test = train = pd.read_csv(out, sep=\\\",\\\", low_memory=False)\";\n",
       "                var nbb_formatted_code = \"kaggle_test = train = pd.read_csv(out, sep=\\\",\\\", low_memory=False)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "kaggle_test = train = pd.read_csv(out, sep=\",\", low_memory=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.347927Z",
     "start_time": "2020-09-19T17:12:39.287807Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 53;\n",
       "                var nbb_unformatted_code = \"input_kaggle, _ = preproc_dataset(kaggle_test, target, [\\\"Set\\\"], params)\";\n",
       "                var nbb_formatted_code = \"input_kaggle, _ = preproc_dataset(kaggle_test, target, [\\\"Set\\\"], params)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "input_kaggle, _ = preproc_dataset(kaggle_test, target, [\"Set\"], params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.662930Z",
     "start_time": "2020-09-19T17:12:39.348884Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 54;\n",
       "                var nbb_unformatted_code = \"probs_kaggle, _ = predict(model, input_kaggle)\";\n",
       "                var nbb_formatted_code = \"probs_kaggle, _ = predict(model, input_kaggle)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "probs_kaggle, _ = predict(model, input_kaggle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.686021Z",
     "start_time": "2020-09-19T17:12:39.663786Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 55;\n",
       "                var nbb_unformatted_code = \"probs_kaggle = target_encoder.inverse_transform(probs_kaggle)\";\n",
       "                var nbb_formatted_code = \"probs_kaggle = target_encoder.inverse_transform(probs_kaggle)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "probs_kaggle = target_encoder.inverse_transform(probs_kaggle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.710077Z",
     "start_time": "2020-09-19T17:12:39.686953Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[118826.086],\n",
       "       [168432.67 ],\n",
       "       [204974.73 ],\n",
       "       ...,\n",
       "       [191878.77 ],\n",
       "       [119177.06 ],\n",
       "       [225938.64 ]], dtype=float32)"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 56;\n",
       "                var nbb_unformatted_code = \"probs_kaggle\";\n",
       "                var nbb_formatted_code = \"probs_kaggle\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "probs_kaggle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.735503Z",
     "start_time": "2020-09-19T17:12:39.711248Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 57;\n",
       "                var nbb_unformatted_code = \"test_df = kaggle_test[[\\\"Id\\\"]]\";\n",
       "                var nbb_formatted_code = \"test_df = kaggle_test[[\\\"Id\\\"]]\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_df = kaggle_test[[\"Id\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.760621Z",
     "start_time": "2020-09-19T17:12:39.736368Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/work/.cache/poetry/xplainet-KS9Ipbc1-py3.7/lib/python3.7/site-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 58;\n",
       "                var nbb_unformatted_code = \"test_df[\\\"SalePrice\\\"] = probs_kaggle\";\n",
       "                var nbb_formatted_code = \"test_df[\\\"SalePrice\\\"] = probs_kaggle\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_df[\"SalePrice\"] = probs_kaggle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.788793Z",
     "start_time": "2020-09-19T17:12:39.761984Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 59;\n",
       "                var nbb_unformatted_code = \"test_df.to_csv(\\\"kaggle_submit.csv\\\", index=False)\";\n",
       "                var nbb_formatted_code = \"test_df.to_csv(\\\"kaggle_submit.csv\\\", index=False)\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_df.to_csv(\"kaggle_submit.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-09-19T17:12:39.813555Z",
     "start_time": "2020-09-19T17:12:39.789893Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "            setTimeout(function() {\n",
       "                var nbb_cell_id = 60;\n",
       "                var nbb_unformatted_code = \"# Scores 0.18047 on kaggle\\n# https://www.kaggle.com/c/house-prices-advanced-regression-techniques/leaderboard#score\\n# Should train with RMSLE ?\\n# from tensorflow.keras import backend as K\\n\\n# def root_mean_squared_log_error(y_true, y_pred):\\n#     return K.sqrt(K.mean(K.square(K.log(1+y_pred) - K.log(1+y_true))))\";\n",
       "                var nbb_formatted_code = \"# Scores 0.18047 on kaggle\\n# https://www.kaggle.com/c/house-prices-advanced-regression-techniques/leaderboard#score\\n# Should train with RMSLE ?\\n# from tensorflow.keras import backend as K\\n\\n# def root_mean_squared_log_error(y_true, y_pred):\\n#     return K.sqrt(K.mean(K.square(K.log(1+y_pred) - K.log(1+y_true))))\";\n",
       "                var nbb_cells = Jupyter.notebook.get_cells();\n",
       "                for (var i = 0; i < nbb_cells.length; ++i) {\n",
       "                    if (nbb_cells[i].input_prompt_number == nbb_cell_id) {\n",
       "                        if (nbb_cells[i].get_text() == nbb_unformatted_code) {\n",
       "                             nbb_cells[i].set_text(nbb_formatted_code);\n",
       "                        }\n",
       "                        break;\n",
       "                    }\n",
       "                }\n",
       "            }, 500);\n",
       "            "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Scores 0.18047 on kaggle\n",
    "# https://www.kaggle.com/c/house-prices-advanced-regression-techniques/leaderboard#score\n",
    "# Should train with RMSLE ?\n",
    "# from tensorflow.keras import backend as K\n",
    "\n",
    "# def root_mean_squared_log_error(y_true, y_pred):\n",
    "#     return K.sqrt(K.mean(K.square(K.log(1+y_pred) - K.log(1+y_true))))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}