plot_feature_selection.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import os\n",
    "import time\n",
    "\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "PH = 60\n",
    "patient_list = [559, 570, 588, 563, 575, 591]\n",
    "patient_list = [540,544,552,567,584,596]\n",
    "# patient_list = [540]\n",
    "\n",
    "df_full = pd.DataFrame()\n",
    "for patN in patient_list:\n",
    "    # Read data\n",
    "    patient = str(patN)\n",
    "    data_tr = pd.read_csv(F'data/ohio{patient}_Training.txt')\n",
    "    # Load training results\n",
    "    out_str = 'single_output'\n",
    "    results = np.load(os.path.join('results','hpar_full_selection','results_training.'+patient+'.ph'+str(PH)+'.'+out_str+'.npy'),allow_pickle='TRUE').item()\n",
    "    df = pd.DataFrame.from_dict(results)\n",
    "    df['to_use'] = [','.join(l) for l in df['to_use']] # converts list to string\n",
    "    df['patN'] = patN\n",
    "    # normalize score with min-max scaling\n",
    "    x = MinMaxScaler().fit_transform(df['val_loss'].values.reshape(-1,1))\n",
    "    df['val_loss'] = x\n",
    "    # concatenate patient results\n",
    "    df_full = pd.concat([df_full, df])\n",
    "\n",
    "df = df_full"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "figsize = [15,5]\n",
    "\n",
    "fig, ax = plt.subplots(1,1,figsize=figsize)\n",
    "sns.boxplot(x='split_prc', y=\"val_loss\", data=df)\n",
    "for tick in ax.get_xticklabels():\n",
    "    tick.set_rotation(90)\n",
    "\n",
    "fig, ax = plt.subplots(1,1,figsize=figsize)\n",
    "my_order = df.groupby(by=[\"to_use\"])[\"val_loss\"].median().sort_values().index\n",
    "sns.boxplot(x='to_use', y=\"val_loss\", data=df, order=my_order)\n",
    "for tick in ax.get_xticklabels():\n",
    "    tick.set_rotation(90)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_style(\"whitegrid\")\n",
    "\n",
    "figsize = [12,5]\n",
    "fig, ax = plt.subplots(1,1,figsize=figsize)\n",
    "df_view = df.sort_values(by=[\"val_loss\"])\n",
    "\n",
    "feat = 'CGM'\n",
    "df_view.loc[:,feat] = [(feat in x) for x in df_view['to_use']]\n",
    "\n",
    "df_view.loc[:,'x'] = np.arange(0,len(df_view['val_loss']),1)\n",
    "sns.scatterplot(x='x', y=\"val_loss\", data=df_view, hue=feat)\n",
    "ax.grid(axis='x')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "feat_importance = {}\n",
    "\n",
    "full_feat_set = ['CGM','DR','pie','pce','ic_pe','exercise_ob_2nd','basis_heart_rate']\n",
    "full_feat_set = ['DR','pie','pce','ic_pe','exercise_ob_2nd']\n",
    "\n",
    "for feat in full_feat_set:\n",
    "    n = 0\n",
    "    for k, el in enumerate(df_view['to_use']):\n",
    "        if feat in el:\n",
    "            n = n + (len(df_view['to_use']) - k)\n",
    "    feat_importance[feat] = n\n",
    "\n",
    "feature_names = [x for x in feat_importance.keys()]\n",
    "\n",
    "def minmaxscaler(x):\n",
    "    return (x - min(x)) / (max(x - min(x)))\n",
    "\n",
    "feat_importance = np.array(list(feat_importance.values()))\n",
    "# foo = minmaxscaler(feat_importance)\n",
    "foo = feat_importance\n",
    "\n",
    "# plot figure\n",
    "fig, ax = plt.subplots(1,1)\n",
    "x = np.arange(0, len(foo), 1)\n",
    "ax.barh(x, foo, 0.65)\n",
    "ax.set_yticks(x)\n",
    "ax.set_yticklabels(feature_names)\n",
    "ax.set_title('Feature ranking (normalized)')\n",
    "ax.grid(b=True, which='major', axis='both')\n",
    "\n",
    "# plot figure\n",
    "fig, ax = plt.subplots(1,1)\n",
    "x = np.arange(0, len(foo), 1)\n",
    "ax.barh(x, minmaxscaler(foo), 0.65)\n",
    "ax.set_yticks(x)\n",
    "ax.set_yticklabels(feature_names)\n",
    "ax.set_title('Feature ranking (minmax scaling)')\n",
    "ax.grid(b=True, which='major', axis='both')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>val_loss</th>\n",
       "      <th>split_prc</th>\n",
       "      <th>patN</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>to_use</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pie</th>\n",
       "      <td>0.301932</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pie,pce</th>\n",
       "      <td>0.320158</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pie</th>\n",
       "      <td>0.322364</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pie,pce</th>\n",
       "      <td>0.341148</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pie,ic_pe</th>\n",
       "      <td>0.361729</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pie,ic_pe</th>\n",
       "      <td>0.363408</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pie,exercise_ob_2nd</th>\n",
       "      <td>0.386937</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pie,exercise_ob_2nd</th>\n",
       "      <td>0.390359</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pce,ic_pe</th>\n",
       "      <td>0.392872</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pce,ic_pe</th>\n",
       "      <td>0.393107</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pie,pce,ic_pe</th>\n",
       "      <td>0.394091</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pie,pce,ic_pe</th>\n",
       "      <td>0.408040</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pie,pce,exercise_ob_2nd</th>\n",
       "      <td>0.411863</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pie,pce,exercise_ob_2nd</th>\n",
       "      <td>0.416596</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pie,ic_pe,exercise_ob_2nd</th>\n",
       "      <td>0.431147</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,ic_pe</th>\n",
       "      <td>0.442286</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,ic_pe</th>\n",
       "      <td>0.444997</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pie,ic_pe,exercise_ob_2nd</th>\n",
       "      <td>0.446456</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pce,ic_pe,exercise_ob_2nd</th>\n",
       "      <td>0.482997</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pie,pce,ic_pe,exercise_ob_2nd</th>\n",
       "      <td>0.484345</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pie,pce,ic_pe,exercise_ob_2nd</th>\n",
       "      <td>0.484404</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pce</th>\n",
       "      <td>0.486264</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pce,ic_pe,exercise_ob_2nd</th>\n",
       "      <td>0.491634</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pce</th>\n",
       "      <td>0.500250</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,ic_pe,exercise_ob_2nd</th>\n",
       "      <td>0.523542</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,ic_pe,exercise_ob_2nd</th>\n",
       "      <td>0.524758</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR</th>\n",
       "      <td>0.538272</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM</th>\n",
       "      <td>0.570131</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,pce,exercise_ob_2nd</th>\n",
       "      <td>0.585479</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,pce,exercise_ob_2nd</th>\n",
       "      <td>0.596376</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,DR,exercise_ob_2nd</th>\n",
       "      <td>0.640362</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGM,exercise_ob_2nd</th>\n",
       "      <td>0.640378</td>\n",
       "      <td>0.65</td>\n",
       "      <td>563.833333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                      val_loss  split_prc        patN\n",
       "to_use                                                               \n",
       "CGM,DR,pie                            0.301932       0.65  563.833333\n",
       "CGM,pie,pce                           0.320158       0.65  563.833333\n",
       "CGM,pie                               0.322364       0.65  563.833333\n",
       "CGM,DR,pie,pce                        0.341148       0.65  563.833333\n",
       "CGM,pie,ic_pe                         0.361729       0.65  563.833333\n",
       "CGM,DR,pie,ic_pe                      0.363408       0.65  563.833333\n",
       "CGM,DR,pie,exercise_ob_2nd            0.386937       0.65  563.833333\n",
       "CGM,pie,exercise_ob_2nd               0.390359       0.65  563.833333\n",
       "CGM,DR,pce,ic_pe                      0.392872       0.65  563.833333\n",
       "CGM,pce,ic_pe                         0.393107       0.65  563.833333\n",
       "CGM,pie,pce,ic_pe                     0.394091       0.65  563.833333\n",
       "CGM,DR,pie,pce,ic_pe                  0.408040       0.65  563.833333\n",
       "CGM,pie,pce,exercise_ob_2nd           0.411863       0.65  563.833333\n",
       "CGM,DR,pie,pce,exercise_ob_2nd        0.416596       0.65  563.833333\n",
       "CGM,pie,ic_pe,exercise_ob_2nd         0.431147       0.65  563.833333\n",
       "CGM,ic_pe                             0.442286       0.65  563.833333\n",
       "CGM,DR,ic_pe                          0.444997       0.65  563.833333\n",
       "CGM,DR,pie,ic_pe,exercise_ob_2nd      0.446456       0.65  563.833333\n",
       "CGM,pce,ic_pe,exercise_ob_2nd         0.482997       0.65  563.833333\n",
       "CGM,DR,pie,pce,ic_pe,exercise_ob_2nd  0.484345       0.65  563.833333\n",
       "CGM,pie,pce,ic_pe,exercise_ob_2nd     0.484404       0.65  563.833333\n",
       "CGM,DR,pce                            0.486264       0.65  563.833333\n",
       "CGM,DR,pce,ic_pe,exercise_ob_2nd      0.491634       0.65  563.833333\n",
       "CGM,pce                               0.500250       0.65  563.833333\n",
       "CGM,ic_pe,exercise_ob_2nd             0.523542       0.65  563.833333\n",
       "CGM,DR,ic_pe,exercise_ob_2nd          0.524758       0.65  563.833333\n",
       "CGM,DR                                0.538272       0.65  563.833333\n",
       "CGM                                   0.570131       0.65  563.833333\n",
       "CGM,DR,pce,exercise_ob_2nd            0.585479       0.65  563.833333\n",
       "CGM,pce,exercise_ob_2nd               0.596376       0.65  563.833333\n",
       "CGM,DR,exercise_ob_2nd                0.640362       0.65  563.833333\n",
       "CGM,exercise_ob_2nd                   0.640378       0.65  563.833333"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('to_use').mean().sort_values('val_loss')"
   ]
  }
 ],
 "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": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}