bidd-aggmap / Git / [03b52a] /paper/NAR/03_COVID-19/01_covid-19

Models:
AlyssaS/
bidd-aggmap
Downloads: 1
[03b52a]: / paper / NAR / 03_COVID-19 / 01_covid-19_C135.ipynb
History
Download this file
1824 lines (1823 with data), 197.0 kB

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAV0AAABECAYAAAAiJuZQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAACU0lEQVR4nO3ZsWoUURiG4T/ZoBtBCMpAIEIsrHTBJngZ3sA23lxuwBuQpEw9XoJgnGZR0bGYHYugQpLpPP+J4/OUexj4FoaX5ezOOI5jAJBit/YAgP+J6AIkEl2ARKILkGhv6qDv+2jbNpqmicVikbkJ4J81DEN0XRer1SqWy+WN88notm0b6/W66DiAuTo9PY2Tk5Mbn09Gt2maqwdfPYzD5TxvITavX9aeUNTZi+e1JxTzLp7VnlDU2fuj2hOKevr2oPaEYrb9ZXy9ePO7oddNRvfXlcLhcjee7M/zemG/eVB7QlEHR/N9se/H7S/0XGw/HdaeUNRi/3HtCcVNXcvO8ycswB0lugCJRBcgkegCJBJdgESiC5BIdAESiS5AItEFSCS6AIlEFyCR6AIkEl2ARKILkEh0ARKJLkAi0QVIJLoAiUQXIJHoAiQSXYBEoguQSHQBEokuQCLRBUgkugCJRBcgkegCJBJdgESiC5BIdAESiS5AItEFSCS6AIlEFyCR6AIkEl2ARKILkEh0ARKJLkAi0QVIJLoAiUQXIJHoAiQSXYBEoguQSHQBEokuQCLRBUgkugCJRBcg0d7UwTAMERHxsd+mjcm26b7VnlDU5sOm9oRifkRXe0JRu9292hOKGr73tScUs+0vI+JPQ6+bjG7XXb3U64svBWbdEefntRcUNvfvN1+Pag8o7HPtAQm6rovj4+Mbn++M4zje9kDf99G2bTRNE4vFovhAgDkYhiG6rovVahXL5fLG+WR0Afj7/JEGkEh0ARKJLkAi0QVI9BOYSVxbe78KfwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x72 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# -*- coding: utf-8 -*-\n",
    "\"\"\"\n",
    "Created on Sun Aug 16 17:10:53 2020\n",
    "\n",
    "@author: wanxiang.shen@u.nus.edu\n",
    "\"\"\"\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from sklearn.model_selection import KFold, StratifiedKFold\n",
    "from sklearn.metrics import confusion_matrix, precision_recall_curve\n",
    "from sklearn.metrics import auc as calculate_auc\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "from aggmap import AggMap, AggModel\n",
    "\n",
    "np.random.seed(666) #just for reaptable results\n",
    "\n",
    "\n",
    "color = sns.color_palette(\"rainbow_r\", 6) #PiYG\n",
    "sns.palplot(color)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:00:30,354 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Calculating distance ...\u001b[0m\n",
      "2020-10-25 15:00:30,354 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - the number of process is 16\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 3828/3828 [00:00<00:00, 12543.09it/s]\n",
      "100%|##########| 3828/3828 [00:00<00:00, 1803414.10it/s]\n",
      "100%|##########| 88/88 [00:00<00:00, 756.39it/s]\n"
     ]
    }
   ],
   "source": [
    "dfx  = pd.read_excel('./COVID19.xlsx', sheet_name='data')\n",
    "dfy  = pd.read_excel('./COVID19.xlsx', sheet_name='sample_info')\n",
    "\n",
    "dfx = dfx[dfx.columns[1:]]\n",
    "cols = [\"p-%s\" % c for c in dfx.columns]\n",
    "dfx.columns = cols\n",
    "\n",
    "X = dfx.values\n",
    "Y = pd.get_dummies(dfy['class']).values.astype(float)\n",
    "\n",
    "mp = AggMap(dfx, metric = 'correlation')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5-fold cross validation and repeat 5 times "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:00:32,254 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:00:32,434 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:00:32,444 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:02<00:00, 102.57it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1592.49it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 1), (73, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1704.37it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1612.12it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 1), (73, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1623.55it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1394.75it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1536.80it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 2003.12it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1586.64it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1500.41it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:02:36,139 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:02:36,321 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:02:36,332 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1269.65it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1659.43it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 3), (73, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1483.53it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1632.17it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 3), (73, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1658.82it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1407.64it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1389.52it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1664.97it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1620.26it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1336.30it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n",
      "2020-10-25 15:04:37,788 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:04:37,976 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1673.56it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:04:37,990 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "100%|##########| 73/73 [00:00<00:00, 1741.36it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 5), (73, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1656.33it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1356.36it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 5), (73, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1653.54it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1795.98it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1789.66it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1750.37it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1666.12it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1203.72it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:06:39,109 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:06:39,288 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:06:39,298 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1646.38it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1593.27it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 1), (73, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1634.97it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1712.09it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 1), (73, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1526.68it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1521.88it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1695.80it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1521.50it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1525.84it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1686.31it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n",
      "2020-10-25 15:08:40,151 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:08:40,349 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 22%|##2       | 64/289 [00:00<00:00, 586.89it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:08:40,363 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1275.42it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1695.32it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 3), (73, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1511.63it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1810.79it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 3), (73, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1615.37it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1762.98it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1637.86it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1682.38it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1523.87it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1741.26it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:10:41,087 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:10:41,267 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:10:41,280 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1654.40it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1416.57it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 5), (73, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1700.28it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1725.68it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 5), (73, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1411.22it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1239.99it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1638.17it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1479.96it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1730.13it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 2042.80it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:12:42,348 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:12:42,528 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:12:42,539 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1490.83it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1839.98it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 1), (73, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1461.25it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1420.97it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 1), (73, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1460.50it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1567.85it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1568.72it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1772.19it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1705.71it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1785.50it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n",
      "2020-10-25 15:14:42,974 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:14:43,164 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:14:43,175 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1810.62it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1285.44it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 3), (73, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1623.81it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1749.49it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 3), (73, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1460.08it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 2040.32it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1600.04it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1685.75it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1703.98it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1010.23it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:16:44,254 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:16:44,435 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:16:44,448 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1602.56it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1648.18it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 5), (73, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1620.29it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1124.99it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 5), (73, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1559.32it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 2160.82it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1646.85it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1296.02it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1604.73it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1512.93it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:18:46,082 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:18:46,264 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:18:46,274 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1401.48it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1869.04it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 1), (73, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1606.90it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1597.45it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 1), (73, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1392.59it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1429.80it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1691.53it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1801.62it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1590.68it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1469.22it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:20:46,594 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:20:46,775 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:20:46,786 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1538.23it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 2184.94it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 3), (73, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1492.92it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1427.80it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 3), (73, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1309.92it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1512.85it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1350.82it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1688.07it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1637.20it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1448.62it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:22:47,977 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:22:48,164 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:22:48,177 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1611.45it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1599.78it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 5), (73, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1570.96it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1498.88it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 5), (73, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1672.30it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1029.90it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1756.66it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1577.68it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1408.17it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1499.01it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:24:49,390 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:24:49,571 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:24:49,580 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1378.69it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 2325.50it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 1), (73, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1530.05it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1644.72it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 1), (73, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1582.60it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1033.10it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1364.80it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1678.73it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1440.55it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1440.97it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 1), (72, 10, 9, 1) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:26:50,276 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:26:50,456 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:26:50,468 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1650.33it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1487.25it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 3), (73, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1614.78it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1698.20it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 3), (73, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1846.41it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 2095.93it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1507.46it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1778.05it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1743.65it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1378.54it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 3), (72, 10, 9, 3) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/289 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-10-25 15:28:49,736 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - applying hierarchical clustering to obtain group information ...\u001b[0m\n",
      "2020-10-25 15:28:49,915 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Applying grid feature map(assignment), this may take several minutes(1~30 min)\u001b[0m\n",
      "2020-10-25 15:28:49,929 - \u001b[32mINFO\u001b[0m - [bidd-aggmap]\u001b[0m - Finished\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1853.43it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1519.70it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 5), (73, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 289/289 [00:00<00:00, 1651.63it/s]\n",
      "100%|##########| 73/73 [00:00<00:00, 1726.99it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (289, 10, 9, 5), (73, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1689.53it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1990.35it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1628.23it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1715.42it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|##########| 290/290 [00:00<00:00, 1568.67it/s]\n",
      "100%|##########| 72/72 [00:00<00:00, 1247.69it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " input train and test X shape is (290, 10, 9, 5), (72, 10, 9, 5) \n",
      "MultiClassEstimator(batch_size=4, epochs=50, gpuid='1')\n"
     ]
    }
   ],
   "source": [
    "n_splits = 5 #5-fold reapeat 5 times\n",
    "\n",
    "run_all = []\n",
    "\n",
    "for repeat_seed in [8, 16, 32, 64, 128]: #5 repeats random seeds\n",
    "    outer = KFold(n_splits = n_splits, shuffle = True, random_state = repeat_seed)\n",
    "    outer_idx = list(outer.split(range(len(Y))))\n",
    "    \n",
    "    for c in [1, 3, 5]: \n",
    "        \n",
    "        mp = mp.fit(cluster_channels = c, verbose = 0)\n",
    "        for i, idx in enumerate(outer_idx):\n",
    "            \n",
    "            train_idx, valid_idx = idx\n",
    "            fold_num = \"fold_%s\" % str(i+1).zfill(2) \n",
    "            \n",
    "            validY = Y[valid_idx]\n",
    "            validx = X[valid_idx]\n",
    "\n",
    "            trainx = X[train_idx]\n",
    "            trainY = Y[train_idx]\n",
    "\n",
    "            trainX = mp.batch_transform(trainx, scale_method = 'standard')\n",
    "            validX = mp.batch_transform(validx, scale_method = 'standard')\n",
    "            print(\"\\n input train and test X shape is %s, %s \" % (trainX.shape,  validX.shape))\n",
    "            clf = AggModel.MultiClassEstimator(epochs = 50,  \n",
    "                                               batch_size = 4,\n",
    "                                               verbose = 0, gpuid=1)\n",
    "            clf.fit(trainX, trainY, X_valid = validX, y_valid = validY)\n",
    "\n",
    "            history = clf.history.history\n",
    "            history['fold'] = fold_num\n",
    "            history['c'] = c\n",
    "            history['repeat_seed'] = repeat_seed\n",
    "            \n",
    "            run_all.append(history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "dfr = pd.DataFrame(run_all)\n",
    "dfr.to_csv('./5FCV_valid.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(style = 'white', font_scale=1.8)\n",
    "\n",
    "fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(18,6), sharex=False, sharey=False, )\n",
    "ax1, ax2 = axes        \n",
    "\n",
    "\n",
    "col = 'val_loss'\n",
    "loss_mean = dfr.groupby(['c']).agg({col: lambda x:x.tolist()})[col].apply(lambda x: np.array(x).mean(axis=0)).apply(pd.Series).T\n",
    "loss_std = dfr.groupby(['c']).agg({col: lambda x:x.tolist()})[col].apply(lambda x: np.array(x).std(axis=0)).apply(pd.Series).T\n",
    "\n",
    "loss_mean.columns.name = 'Channels'\n",
    "loss_std.columns.name = 'Channels'\n",
    "\n",
    "loss_mean.plot(ax=ax1, lw = 3, color = color)\n",
    "ax1.set_ylabel('Validation loss')\n",
    "ax1.set_xlabel('Epochs')\n",
    "\n",
    "# for i, j in zip(loss_mean.columns, color[:3]):\n",
    "#     bottom = loss_mean[i]-loss_std[i]\n",
    "#     up = loss_mean[i] + loss_std[i]\n",
    "#     ax1.fill_between(loss_mean.index, bottom, up,alpha=0.2, color = j)\n",
    "\n",
    "#plt.savefig(os.path.join(results_save_folder, 'performance_01_test.png'), bbox_inches='tight', dpi=300)\n",
    "\n",
    "col = 'val_accuracy'\n",
    "acc_mean = dfr.groupby(['c']).agg({col: lambda x:x.tolist()})[col].apply(lambda x: np.array(x).mean(axis=0)).apply(pd.Series).T\n",
    "acc_std = dfr.groupby(['c']).agg({col: lambda x:x.tolist()})[col].apply(lambda x: np.array(x).std(axis=0)).apply(pd.Series).T\n",
    "\n",
    "\n",
    "acc_mean.columns.name = 'Channels'\n",
    "acc_std.columns.name = 'Channels'\n",
    "\n",
    "acc_mean.plot(ax=ax2, lw = 3, color = color)\n",
    "ax2.set_ylabel('Validation accuracy')\n",
    "ax2.set_xlabel('Epochs')\n",
    "\n",
    "fig.tight_layout()\n",
    "plt.subplots_adjust(left=None, bottom=None, right=None, top=None, wspace=0.3)\n",
    "plt.savefig('5FCV_valid.png', bbox_inches='tight', dpi=400)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "dfr = pd.read_csv('./5FCV_valid.csv',index_col=0)\n",
    "f = lambda x:[float(i) for i in x.replace(\"[\",'').replace(\"]\",'').split(',')]\n",
    "dfr['accuracy'] = dfr['accuracy'].apply(f)\n",
    "dfr['val_accuracy'] = dfr['val_accuracy'].apply(f)\n",
    "dfr['val_loss'] = dfr['val_loss'].apply(f)\n",
    "dfr['loss'] = dfr['loss'].apply(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(style = 'white', font_scale=2)\n",
    "\n",
    "fig, axes = plt.subplots(nrows=1, ncols=1, figsize=(9,6), sharex=False, sharey=False, )\n",
    "ax2 = axes        \n",
    "\n",
    "col = 'val_accuracy'\n",
    "acc_mean = dfr.groupby(['c']).agg({col: lambda x:x.tolist()})[col].apply(lambda x: np.array(x).mean(axis=0)).apply(pd.Series).T\n",
    "acc_std = dfr.groupby(['c']).agg({col: lambda x:x.tolist()})[col].apply(lambda x: np.array(x).std(axis=0)).apply(pd.Series).T\n",
    "\n",
    "\n",
    "acc_mean.columns.name = 'Channels'\n",
    "acc_std.columns.name = 'Channels'\n",
    "\n",
    "acc_mean.plot(ax=ax2, lw = 4, color = color)\n",
    "ax2.set_ylabel('Accuracy')\n",
    "ax2.set_xlabel('Epochs')\n",
    "#plt.title('COVMS')\n",
    "\n",
    "ax2.tick_params(bottom='on', left='off',  labelleft='on', labelbottom='on', pad=-.6,)\n",
    "sns.despine(top=True, right=True, left=False, bottom=False)\n",
    "\n",
    "fig.tight_layout()\n",
    "plt.subplots_adjust(left=None, bottom=None, right=None, top=None, wspace=0.3)\n",
    "#plt.savefig('5FCV_valid_acc.pdf', bbox_inches='tight', dpi=400)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}