diff --git a/clinical_model/clinical_data.ipynb b/clinical_model/clinical_data.ipynb
new file mode 100644
index 0000000..210dbd1
--- /dev/null
+++ b/clinical_model/clinical_data.ipynb
@@ -0,0 +1,2154 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "electoral-korean",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "import pathlib\n",
+    "import numpy as np"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "boolean-bacteria",
+   "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>Age</th>\n",
+       "      <th>Sex</th>\n",
+       "      <th>Temp_C</th>\n",
+       "      <th>Cough</th>\n",
+       "      <th>DifficultyInBreathing</th>\n",
+       "      <th>WBC</th>\n",
+       "      <th>CRP</th>\n",
+       "      <th>Fibrinogen</th>\n",
+       "      <th>LDH</th>\n",
+       "      <th>Ddimer</th>\n",
+       "      <th>Ox_percentage</th>\n",
+       "      <th>PaO2</th>\n",
+       "      <th>SaO2</th>\n",
+       "      <th>pH</th>\n",
+       "      <th>CardiovascularDisease</th>\n",
+       "      <th>RespiratoryFailure</th>\n",
+       "      <th>Prognosis</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>PatientID</th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>P_131</th>\n",
+       "      <td>35.913889</td>\n",
+       "      <td>0</td>\n",
+       "      <td>39.3</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.76</td>\n",
+       "      <td>43.40</td>\n",
+       "      <td>651.0</td>\n",
+       "      <td>387.0</td>\n",
+       "      <td>157.0</td>\n",
+       "      <td>94.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_132</th>\n",
+       "      <td>57.266667</td>\n",
+       "      <td>0</td>\n",
+       "      <td>37.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>11.48</td>\n",
+       "      <td>64.00</td>\n",
+       "      <td>700.0</td>\n",
+       "      <td>338.0</td>\n",
+       "      <td>601.0</td>\n",
+       "      <td>94.0</td>\n",
+       "      <td>75.0</td>\n",
+       "      <td>96.9</td>\n",
+       "      <td>7.42</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_195</th>\n",
+       "      <td>79.263889</td>\n",
+       "      <td>0</td>\n",
+       "      <td>37.8</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.21</td>\n",
+       "      <td>115.30</td>\n",
+       "      <td>698.0</td>\n",
+       "      <td>356.0</td>\n",
+       "      <td>448.0</td>\n",
+       "      <td>94.0</td>\n",
+       "      <td>63.0</td>\n",
+       "      <td>94.6</td>\n",
+       "      <td>7.39</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_193</th>\n",
+       "      <td>82.000000</td>\n",
+       "      <td>0</td>\n",
+       "      <td>38.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>7.28</td>\n",
+       "      <td>149.30</td>\n",
+       "      <td>513.0</td>\n",
+       "      <td>482.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>97.0</td>\n",
+       "      <td>68.0</td>\n",
+       "      <td>96.3</td>\n",
+       "      <td>7.46</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_140</th>\n",
+       "      <td>60.791667</td>\n",
+       "      <td>1</td>\n",
+       "      <td>37.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.37</td>\n",
+       "      <td>20.70</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>210.0</td>\n",
+       "      <td>93.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>97.3</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_12</th>\n",
+       "      <td>51.000000</td>\n",
+       "      <td>0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>14.30</td>\n",
+       "      <td>22.79</td>\n",
+       "      <td>550.0</td>\n",
+       "      <td>368.0</td>\n",
+       "      <td>5027.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_8</th>\n",
+       "      <td>57.000000</td>\n",
+       "      <td>0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.10</td>\n",
+       "      <td>8.93</td>\n",
+       "      <td>757.0</td>\n",
+       "      <td>451.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_10</th>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>7.30</td>\n",
+       "      <td>0.23</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>1073.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_26</th>\n",
+       "      <td>92.000000</td>\n",
+       "      <td>1</td>\n",
+       "      <td>38.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>13.50</td>\n",
+       "      <td>3.77</td>\n",
+       "      <td>533.0</td>\n",
+       "      <td>358.0</td>\n",
+       "      <td>2154.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>75.7</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>7.36</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_146</th>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>0</td>\n",
+       "      <td>38.4</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.70</td>\n",
+       "      <td>0.02</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>149.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>80.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>7.46</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>863 rows × 17 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                 Age  Sex  Temp_C  Cough  DifficultyInBreathing    WBC  \\\n",
+       "PatientID                                                                \n",
+       "P_131      35.913889    0    39.3    1.0                    0.0   5.76   \n",
+       "P_132      57.266667    0    37.0    0.0                    0.0  11.48   \n",
+       "P_195      79.263889    0    37.8    1.0                    0.0   6.21   \n",
+       "P_193      82.000000    0    38.0    1.0                    0.0   7.28   \n",
+       "P_140      60.791667    1    37.0    1.0                    0.0   6.37   \n",
+       "...              ...  ...     ...    ...                    ...    ...   \n",
+       "P_1_12     51.000000    0     NaN    0.0                    1.0  14.30   \n",
+       "P_1_8      57.000000    0     NaN    1.0                    0.0   5.10   \n",
+       "P_1_10     38.000000    0     NaN    0.0                    1.0   7.30   \n",
+       "P_1_26     92.000000    1    38.0    0.0                    1.0  13.50   \n",
+       "P_1_146    80.000000    0    38.4    0.0                    0.0   6.70   \n",
+       "\n",
+       "              CRP  Fibrinogen    LDH  Ddimer  Ox_percentage  PaO2  SaO2    pH  \\\n",
+       "PatientID                                                                       \n",
+       "P_131       43.40       651.0  387.0   157.0           94.0   NaN   NaN   NaN   \n",
+       "P_132       64.00       700.0  338.0   601.0           94.0  75.0  96.9  7.42   \n",
+       "P_195      115.30       698.0  356.0   448.0           94.0  63.0  94.6  7.39   \n",
+       "P_193      149.30       513.0  482.0     NaN           97.0  68.0  96.3  7.46   \n",
+       "P_140       20.70         NaN    NaN   210.0           93.0   NaN  97.3   NaN   \n",
+       "...           ...         ...    ...     ...            ...   ...   ...   ...   \n",
+       "P_1_12      22.79       550.0  368.0  5027.0            NaN   NaN   NaN   NaN   \n",
+       "P_1_8        8.93       757.0  451.0     NaN            NaN   NaN   NaN   NaN   \n",
+       "P_1_10       0.23         NaN    NaN  1073.0            NaN   NaN   NaN   NaN   \n",
+       "P_1_26       3.77       533.0  358.0  2154.0            NaN  75.7   NaN  7.36   \n",
+       "P_1_146      0.02         NaN  149.0     NaN            NaN  80.0   NaN  7.46   \n",
+       "\n",
+       "           CardiovascularDisease  RespiratoryFailure  Prognosis  \n",
+       "PatientID                                                        \n",
+       "P_131                        0.0                 NaN          0  \n",
+       "P_132                        0.0                 NaN          0  \n",
+       "P_195                        1.0                 NaN          1  \n",
+       "P_193                        0.0                 NaN          1  \n",
+       "P_140                        0.0                 NaN          0  \n",
+       "...                          ...                 ...        ...  \n",
+       "P_1_12                       0.0                 0.0          1  \n",
+       "P_1_8                        0.0                 0.0          1  \n",
+       "P_1_10                       0.0                 0.0          0  \n",
+       "P_1_26                       0.0                 0.0          1  \n",
+       "P_1_146                      1.0                 0.0          0  \n",
+       "\n",
+       "[863 rows x 17 columns]"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "train_path = pathlib.Path(\"/home/starke88/data/haicu/covid_data_challenge_2021/trainSet.txt\")\n",
+    "df_train = pd.read_csv(train_path)\n",
+    "df_train.set_index(\"PatientID\", inplace=True)\n",
+    "df_train.drop([\"ImageFile\", \"Hospital\"], axis=1, inplace=True)\n",
+    "df_train[\"Prognosis\"] = (df_train[\"Prognosis\"] == \"SEVERE\").astype(np.uint8)\n",
+    "\n",
+    "# one-hot encode the hospital\n",
+    "#foo = pd.get_dummies(df_train[\"Hospital\"], prefix=\"Hospital\")\n",
+    "#df_train = pd.concat([df_train, foo], axis=1)\n",
+    "#df_train = df_train.drop([\"Hospital\"], axis=1)\n",
+    "\n",
+    "\n",
+    "df_train"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "fantastic-struggle",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(6, 17)"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# this is how many complete samples we have\n",
+    "df_train.dropna().shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "eed10a30",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "ongoing-dance",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Age \t\t\t 1 / 863\n",
+      "Sex \t\t\t 0 / 863\n",
+      "Temp_C \t\t\t 154 / 863\n",
+      "Cough \t\t\t 5 / 863\n",
+      "DifficultyInBreathing \t\t\t 4 / 863\n",
+      "WBC \t\t\t 9 / 863\n",
+      "CRP \t\t\t 33 / 863\n",
+      "Fibrinogen \t\t\t 591 / 863\n",
+      "LDH \t\t\t 136 / 863\n",
+      "Ddimer \t\t\t 621 / 863\n",
+      "Ox_percentage \t\t\t 243 / 863\n",
+      "PaO2 \t\t\t 170 / 863\n",
+      "SaO2 \t\t\t 583 / 863\n",
+      "pH \t\t\t 207 / 863\n",
+      "CardiovascularDisease \t\t\t 19 / 863\n",
+      "RespiratoryFailure \t\t\t 159 / 863\n",
+      "Prognosis \t\t\t 0 / 863\n",
+      "\n",
+      "Going to discard features with more than 200 nans: ['Fibrinogen', 'Ddimer', 'Ox_percentage', 'SaO2', 'pH']\n"
+     ]
+    }
+   ],
+   "source": [
+    "# how many samples are missing for individual columns\n",
+    "discard = []\n",
+    "discard_thresh = 200\n",
+    "for c in df_train.columns:\n",
+    "    n_missing = df_train[c].isna().sum()\n",
+    "    print(c, \"\\t\\t\\t\", n_missing, \"/\", len(df_train))\n",
+    "    if n_missing > discard_thresh:\n",
+    "        discard.append(c)\n",
+    "        \n",
+    "print(f\"\\nGoing to discard features with more than {discard_thresh} nans:\", discard)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "induced-treat",
+   "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>Age</th>\n",
+       "      <th>Sex</th>\n",
+       "      <th>Temp_C</th>\n",
+       "      <th>Cough</th>\n",
+       "      <th>DifficultyInBreathing</th>\n",
+       "      <th>WBC</th>\n",
+       "      <th>CRP</th>\n",
+       "      <th>LDH</th>\n",
+       "      <th>PaO2</th>\n",
+       "      <th>CardiovascularDisease</th>\n",
+       "      <th>RespiratoryFailure</th>\n",
+       "      <th>Prognosis</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>PatientID</th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>P_131</th>\n",
+       "      <td>35.913889</td>\n",
+       "      <td>0</td>\n",
+       "      <td>39.3</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.76</td>\n",
+       "      <td>43.40</td>\n",
+       "      <td>387.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_132</th>\n",
+       "      <td>57.266667</td>\n",
+       "      <td>0</td>\n",
+       "      <td>37.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>11.48</td>\n",
+       "      <td>64.00</td>\n",
+       "      <td>338.0</td>\n",
+       "      <td>75.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_195</th>\n",
+       "      <td>79.263889</td>\n",
+       "      <td>0</td>\n",
+       "      <td>37.8</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.21</td>\n",
+       "      <td>115.30</td>\n",
+       "      <td>356.0</td>\n",
+       "      <td>63.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_193</th>\n",
+       "      <td>82.000000</td>\n",
+       "      <td>0</td>\n",
+       "      <td>38.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>7.28</td>\n",
+       "      <td>149.30</td>\n",
+       "      <td>482.0</td>\n",
+       "      <td>68.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_140</th>\n",
+       "      <td>60.791667</td>\n",
+       "      <td>1</td>\n",
+       "      <td>37.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.37</td>\n",
+       "      <td>20.70</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_12</th>\n",
+       "      <td>51.000000</td>\n",
+       "      <td>0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>14.30</td>\n",
+       "      <td>22.79</td>\n",
+       "      <td>368.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_8</th>\n",
+       "      <td>57.000000</td>\n",
+       "      <td>0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.10</td>\n",
+       "      <td>8.93</td>\n",
+       "      <td>451.0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_10</th>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>0</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>7.30</td>\n",
+       "      <td>0.23</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_26</th>\n",
+       "      <td>92.000000</td>\n",
+       "      <td>1</td>\n",
+       "      <td>38.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>13.50</td>\n",
+       "      <td>3.77</td>\n",
+       "      <td>358.0</td>\n",
+       "      <td>75.7</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_146</th>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>0</td>\n",
+       "      <td>38.4</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.70</td>\n",
+       "      <td>0.02</td>\n",
+       "      <td>149.0</td>\n",
+       "      <td>80.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>863 rows × 12 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                 Age  Sex  Temp_C  Cough  DifficultyInBreathing    WBC  \\\n",
+       "PatientID                                                                \n",
+       "P_131      35.913889    0    39.3    1.0                    0.0   5.76   \n",
+       "P_132      57.266667    0    37.0    0.0                    0.0  11.48   \n",
+       "P_195      79.263889    0    37.8    1.0                    0.0   6.21   \n",
+       "P_193      82.000000    0    38.0    1.0                    0.0   7.28   \n",
+       "P_140      60.791667    1    37.0    1.0                    0.0   6.37   \n",
+       "...              ...  ...     ...    ...                    ...    ...   \n",
+       "P_1_12     51.000000    0     NaN    0.0                    1.0  14.30   \n",
+       "P_1_8      57.000000    0     NaN    1.0                    0.0   5.10   \n",
+       "P_1_10     38.000000    0     NaN    0.0                    1.0   7.30   \n",
+       "P_1_26     92.000000    1    38.0    0.0                    1.0  13.50   \n",
+       "P_1_146    80.000000    0    38.4    0.0                    0.0   6.70   \n",
+       "\n",
+       "              CRP    LDH  PaO2  CardiovascularDisease  RespiratoryFailure  \\\n",
+       "PatientID                                                                   \n",
+       "P_131       43.40  387.0   NaN                    0.0                 NaN   \n",
+       "P_132       64.00  338.0  75.0                    0.0                 NaN   \n",
+       "P_195      115.30  356.0  63.0                    1.0                 NaN   \n",
+       "P_193      149.30  482.0  68.0                    0.0                 NaN   \n",
+       "P_140       20.70    NaN   NaN                    0.0                 NaN   \n",
+       "...           ...    ...   ...                    ...                 ...   \n",
+       "P_1_12      22.79  368.0   NaN                    0.0                 0.0   \n",
+       "P_1_8        8.93  451.0   NaN                    0.0                 0.0   \n",
+       "P_1_10       0.23    NaN   NaN                    0.0                 0.0   \n",
+       "P_1_26       3.77  358.0  75.7                    0.0                 0.0   \n",
+       "P_1_146      0.02  149.0  80.0                    1.0                 0.0   \n",
+       "\n",
+       "           Prognosis  \n",
+       "PatientID             \n",
+       "P_131              0  \n",
+       "P_132              0  \n",
+       "P_195              1  \n",
+       "P_193              1  \n",
+       "P_140              0  \n",
+       "...              ...  \n",
+       "P_1_12             1  \n",
+       "P_1_8              1  \n",
+       "P_1_10             0  \n",
+       "P_1_26             1  \n",
+       "P_1_146            0  \n",
+       "\n",
+       "[863 rows x 12 columns]"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_reduced = df_train.drop(discard, axis=1)\n",
+    "df_reduced"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "bizarre-college",
+   "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>Age</th>\n",
+       "      <th>Sex</th>\n",
+       "      <th>Temp_C</th>\n",
+       "      <th>Cough</th>\n",
+       "      <th>DifficultyInBreathing</th>\n",
+       "      <th>WBC</th>\n",
+       "      <th>CRP</th>\n",
+       "      <th>LDH</th>\n",
+       "      <th>PaO2</th>\n",
+       "      <th>CardiovascularDisease</th>\n",
+       "      <th>RespiratoryFailure</th>\n",
+       "      <th>Prognosis</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>count</th>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "      <td>417.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>mean</th>\n",
+       "      <td>65.167866</td>\n",
+       "      <td>0.338129</td>\n",
+       "      <td>37.553717</td>\n",
+       "      <td>0.522782</td>\n",
+       "      <td>0.489209</td>\n",
+       "      <td>6.953189</td>\n",
+       "      <td>32.663094</td>\n",
+       "      <td>370.949640</td>\n",
+       "      <td>73.613669</td>\n",
+       "      <td>0.294964</td>\n",
+       "      <td>0.014388</td>\n",
+       "      <td>0.482014</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>std</th>\n",
+       "      <td>14.489739</td>\n",
+       "      <td>0.473641</td>\n",
+       "      <td>0.968887</td>\n",
+       "      <td>0.500081</td>\n",
+       "      <td>0.500484</td>\n",
+       "      <td>3.279598</td>\n",
+       "      <td>54.263935</td>\n",
+       "      <td>229.063855</td>\n",
+       "      <td>26.501603</td>\n",
+       "      <td>0.456574</td>\n",
+       "      <td>0.119229</td>\n",
+       "      <td>0.500277</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>min</th>\n",
+       "      <td>22.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>35.400000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.200000</td>\n",
+       "      <td>0.010000</td>\n",
+       "      <td>115.000000</td>\n",
+       "      <td>23.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>25%</th>\n",
+       "      <td>56.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>36.800000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>4.700000</td>\n",
+       "      <td>5.360000</td>\n",
+       "      <td>247.000000</td>\n",
+       "      <td>60.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>50%</th>\n",
+       "      <td>65.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>37.700000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>6.300000</td>\n",
+       "      <td>14.420000</td>\n",
+       "      <td>318.000000</td>\n",
+       "      <td>70.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>75%</th>\n",
+       "      <td>78.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>38.200000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>8.460000</td>\n",
+       "      <td>33.960000</td>\n",
+       "      <td>422.000000</td>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>max</th>\n",
+       "      <td>96.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>40.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>26.500000</td>\n",
+       "      <td>413.000000</td>\n",
+       "      <td>2578.000000</td>\n",
+       "      <td>285.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "              Age         Sex      Temp_C       Cough  DifficultyInBreathing  \\\n",
+       "count  417.000000  417.000000  417.000000  417.000000             417.000000   \n",
+       "mean    65.167866    0.338129   37.553717    0.522782               0.489209   \n",
+       "std     14.489739    0.473641    0.968887    0.500081               0.500484   \n",
+       "min     22.000000    0.000000   35.400000    0.000000               0.000000   \n",
+       "25%     56.000000    0.000000   36.800000    0.000000               0.000000   \n",
+       "50%     65.000000    0.000000   37.700000    1.000000               0.000000   \n",
+       "75%     78.000000    1.000000   38.200000    1.000000               1.000000   \n",
+       "max     96.000000    1.000000   40.000000    1.000000               1.000000   \n",
+       "\n",
+       "              WBC         CRP          LDH        PaO2  CardiovascularDisease  \\\n",
+       "count  417.000000  417.000000   417.000000  417.000000             417.000000   \n",
+       "mean     6.953189   32.663094   370.949640   73.613669               0.294964   \n",
+       "std      3.279598   54.263935   229.063855   26.501603               0.456574   \n",
+       "min      0.200000    0.010000   115.000000   23.000000               0.000000   \n",
+       "25%      4.700000    5.360000   247.000000   60.000000               0.000000   \n",
+       "50%      6.300000   14.420000   318.000000   70.000000               0.000000   \n",
+       "75%      8.460000   33.960000   422.000000   80.000000               1.000000   \n",
+       "max     26.500000  413.000000  2578.000000  285.000000               1.000000   \n",
+       "\n",
+       "       RespiratoryFailure   Prognosis  \n",
+       "count          417.000000  417.000000  \n",
+       "mean             0.014388    0.482014  \n",
+       "std              0.119229    0.500277  \n",
+       "min              0.000000    0.000000  \n",
+       "25%              0.000000    0.000000  \n",
+       "50%              0.000000    0.000000  \n",
+       "75%              0.000000    1.000000  \n",
+       "max              1.000000    1.000000  "
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_reduced = df_reduced.dropna()\n",
+    "df_reduced.describe()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "compact-lawsuit",
+   "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>Age</th>\n",
+       "      <th>Sex</th>\n",
+       "      <th>Temp_C</th>\n",
+       "      <th>Cough</th>\n",
+       "      <th>DifficultyInBreathing</th>\n",
+       "      <th>WBC</th>\n",
+       "      <th>CRP</th>\n",
+       "      <th>LDH</th>\n",
+       "      <th>PaO2</th>\n",
+       "      <th>CardiovascularDisease</th>\n",
+       "      <th>RespiratoryFailure</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>PatientID</th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>P_302</th>\n",
+       "      <td>56.0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>38.3</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>7.85</td>\n",
+       "      <td>9.00</td>\n",
+       "      <td>159.0</td>\n",
+       "      <td>73.2</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_301</th>\n",
+       "      <td>61.0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>38.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>3.57</td>\n",
+       "      <td>57.40</td>\n",
+       "      <td>309.0</td>\n",
+       "      <td>57.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_268</th>\n",
+       "      <td>64.0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>36.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.26</td>\n",
+       "      <td>41.90</td>\n",
+       "      <td>299.0</td>\n",
+       "      <td>72.2</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_282</th>\n",
+       "      <td>54.0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>36.6</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>7.01</td>\n",
+       "      <td>67.90</td>\n",
+       "      <td>458.0</td>\n",
+       "      <td>61.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_300</th>\n",
+       "      <td>44.0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>37.7</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>9.18</td>\n",
+       "      <td>42.70</td>\n",
+       "      <td>243.0</td>\n",
+       "      <td>23.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_77</th>\n",
+       "      <td>61.0</td>\n",
+       "      <td>1</td>\n",
+       "      <td>36.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>3.60</td>\n",
+       "      <td>0.07</td>\n",
+       "      <td>219.0</td>\n",
+       "      <td>86.2</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_110</th>\n",
+       "      <td>56.0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>36.5</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>7.00</td>\n",
+       "      <td>13.51</td>\n",
+       "      <td>411.0</td>\n",
+       "      <td>75.8</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_126</th>\n",
+       "      <td>57.0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>37.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>4.75</td>\n",
+       "      <td>2.47</td>\n",
+       "      <td>383.0</td>\n",
+       "      <td>61.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_26</th>\n",
+       "      <td>92.0</td>\n",
+       "      <td>1</td>\n",
+       "      <td>38.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>13.50</td>\n",
+       "      <td>3.77</td>\n",
+       "      <td>358.0</td>\n",
+       "      <td>75.7</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_146</th>\n",
+       "      <td>80.0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>38.4</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.70</td>\n",
+       "      <td>0.02</td>\n",
+       "      <td>149.0</td>\n",
+       "      <td>80.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>417 rows × 11 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "            Age  Sex  Temp_C  Cough  DifficultyInBreathing    WBC    CRP  \\\n",
+       "PatientID                                                                  \n",
+       "P_302      56.0    0    38.3    0.0                    1.0   7.85   9.00   \n",
+       "P_301      61.0    0    38.0    1.0                    1.0   3.57  57.40   \n",
+       "P_268      64.0    0    36.0    1.0                    0.0   5.26  41.90   \n",
+       "P_282      54.0    0    36.6    1.0                    0.0   7.01  67.90   \n",
+       "P_300      44.0    0    37.7    0.0                    0.0   9.18  42.70   \n",
+       "...         ...  ...     ...    ...                    ...    ...    ...   \n",
+       "P_1_77     61.0    1    36.0    1.0                    0.0   3.60   0.07   \n",
+       "P_1_110    56.0    0    36.5    1.0                    0.0   7.00  13.51   \n",
+       "P_1_126    57.0    0    37.0    1.0                    1.0   4.75   2.47   \n",
+       "P_1_26     92.0    1    38.0    0.0                    1.0  13.50   3.77   \n",
+       "P_1_146    80.0    0    38.4    0.0                    0.0   6.70   0.02   \n",
+       "\n",
+       "             LDH  PaO2  CardiovascularDisease  RespiratoryFailure  \n",
+       "PatientID                                                          \n",
+       "P_302      159.0  73.2                    0.0                 0.0  \n",
+       "P_301      309.0  57.0                    0.0                 0.0  \n",
+       "P_268      299.0  72.2                    1.0                 0.0  \n",
+       "P_282      458.0  61.0                    0.0                 0.0  \n",
+       "P_300      243.0  23.0                    0.0                 0.0  \n",
+       "...          ...   ...                    ...                 ...  \n",
+       "P_1_77     219.0  86.2                    0.0                 0.0  \n",
+       "P_1_110    411.0  75.8                    0.0                 0.0  \n",
+       "P_1_126    383.0  61.0                    0.0                 0.0  \n",
+       "P_1_26     358.0  75.7                    0.0                 0.0  \n",
+       "P_1_146    149.0  80.0                    1.0                 0.0  \n",
+       "\n",
+       "[417 rows x 11 columns]"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "label_col = \"Prognosis\"\n",
+    "\n",
+    "y = df_reduced[[label_col]]\n",
+    "X_df = df_reduced[[c for c in df_reduced.columns if c not in [label_col]]]\n",
+    "\n",
+    "X_df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "hollow-match",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "picked_features = X_df.columns\n",
+    "\n",
+    "#picked_features = [\n",
+    "#    \"LDH\",\n",
+    "#    \"Age\",\n",
+    "#    \"PaO2\",\n",
+    "#    \"CRP\",\n",
+    "#    \"WBC\",\n",
+    "#    \"pH\"\n",
+    "#]\n",
+    "\n",
+    "\n",
+    "X = X_df[picked_features]\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "26719885",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# now also load the test data for which maksim imputed the values\n",
+    "\n",
+    "test_path = pathlib.Path(\"/home/starke88/data/haicu/covid_data_challenge_2021/testSet_filled2.txt\")\n",
+    "df_test = pd.read_csv(test_path)\n",
+    "df_test.set_index(\"PatientID\", inplace=True)\n",
+    "#df_test.drop([\"Unnamed: 0\"], axis=1, inplace=True)\n",
+    "# get rid of the features we dumped from the training set as well\n",
+    "df_test.drop([c for c in df_test.columns if c not in df_reduced.columns], axis=1, inplace=True)\n",
+    "# one-hot encode the hospital\n",
+    "#foo = pd.get_dummies(df_test[\"Hospital\"], prefix=\"Hospital\")\n",
+    "#df_test = pd.concat([df_test, foo], axis=1)\n",
+    "#df_test = df_test.drop([\"Hospital\"], axis=1)\n",
+    "assert sorted(df_test.columns) == sorted(X.columns)\n",
+    "\n",
+    "X_test = df_test"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "id": "3cec65d5",
+   "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>Prognosis</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>PatientID</th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>P_102.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_117.png</th>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_16.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_118.png</th>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_114.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_88.png</th>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_92.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_86.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_9.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_90.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>120 rows × 1 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "           Prognosis\n",
+       "PatientID           \n",
+       "P_102.png          1\n",
+       "P_117.png          0\n",
+       "P_16.png           1\n",
+       "P_118.png          0\n",
+       "P_114.png          1\n",
+       "...              ...\n",
+       "P_88.png           0\n",
+       "P_92.png           1\n",
+       "P_86.png           1\n",
+       "P_9.png            1\n",
+       "P_90.png           1\n",
+       "\n",
+       "[120 rows x 1 columns]"
+      ]
+     },
+     "execution_count": 26,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "y_test = pd.read_csv(pathlib.Path(\"/home/starke88/data/haicu/covid_data_challenge_2021/solutionTestSet.txt\"))\n",
+    "y_test.set_index(\"ImageFile\", inplace=True)\n",
+    "y_test.index.rename(\"PatientID\", inplace=True)\n",
+    "y_test[\"Prognosis\"] = (y_test[\"Prognosis\"] == \"SEVERE\").astype(np.uint8)\n",
+    "y_test = y_test[[\"Prognosis\"]]\n",
+    "\n",
+    "y_test"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "placed-ratio",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn.preprocessing import MinMaxScaler, StandardScaler"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "applicable-white",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "312 105\n"
+     ]
+    }
+   ],
+   "source": [
+    "train, valid = train_test_split(X_df.index.to_numpy(), test_size=.25, stratify=y[label_col], random_state=52)\n",
+    "print(len(train), len(valid))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "f230c6c9",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "['P_384' 'P_417' 'P_551' 'P_479' 'P_555' 'P_376' 'P_1_26' 'P_379' 'P_441'\n",
+      " 'P_450' 'P_760' 'P_756' 'P_435' 'P_442' 'P_766' 'P_695' 'P_537' 'P_687'\n",
+      " 'P_467' 'P_1_126' 'P_1_75' 'P_653' 'P_1_149' 'P_411' 'P_358' 'P_740'\n",
+      " 'P_1_81' 'P_634' 'P_616' 'P_382' 'P_342' 'P_519' 'P_476' 'P_639' 'P_759'\n",
+      " 'P_350' 'P_484' 'P_635' 'P_507' 'P_439' 'P_798' 'P_598' 'P_414' 'P_583'\n",
+      " 'P_590' 'P_485' 'P_511' 'P_426' 'P_304' 'P_842' 'P_797' 'P_837' 'P_314'\n",
+      " 'P_572' 'P_641' 'P_803' 'P_1_55' 'P_702' 'P_691' 'P_518' 'P_1_95' 'P_521'\n",
+      " 'P_589' 'P_840' 'P_1_131' 'P_772' 'P_368' 'P_292' 'P_750' 'P_274'\n",
+      " 'P_1_146' 'P_650' 'P_505' 'P_733' 'P_623' 'P_697' 'P_620' 'P_1_140'\n",
+      " 'P_609' 'P_808' 'P_526' 'P_383' 'P_595' 'P_587' 'P_580' 'P_502' 'P_1_22'\n",
+      " 'P_324' 'P_1_72' 'P_725' 'P_336' 'P_491' 'P_781' 'P_407' 'P_546' 'P_645'\n",
+      " 'P_425' 'P_735' 'P_445' 'P_812' 'P_385' 'P_1_66' 'P_743' 'P_560' 'P_633']\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(valid)\n",
+    "pd.DataFrame(valid).to_csv(\"/home/starke88/git/covid_data_challenge/clinical_model/valid_ids.csv\", index=False, header=None)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "advisory-insight",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "array(['P_428', 'P_399', 'P_499', 'P_282'], dtype=object)"
+      ]
+     },
+     "execution_count": 22,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "train[:4]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "convertible-function",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "(312, 11) (312, 1)\n",
+      "(105, 11) (105, 1)\n"
+     ]
+    }
+   ],
+   "source": [
+    "mask_train = X.index.isin(train)\n",
+    "feat_names = X.columns\n",
+    "\n",
+    "X_train = X[mask_train]\n",
+    "y_train = y[mask_train]\n",
+    "\n",
+    "X_valid = X[~mask_train]\n",
+    "y_valid = y[~mask_train]\n",
+    "\n",
+    "print(X_train.shape, y_train.shape)\n",
+    "print(X_valid.shape, y_valid.shape)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "id": "posted-mailing",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "Age                      float64\n",
+       "Sex                        int64\n",
+       "Temp_C                   float64\n",
+       "Cough                    float64\n",
+       "DifficultyInBreathing    float64\n",
+       "WBC                      float64\n",
+       "CRP                      float64\n",
+       "LDH                      float64\n",
+       "PaO2                     float64\n",
+       "CardiovascularDisease    float64\n",
+       "RespiratoryFailure       float64\n",
+       "dtype: object"
+      ]
+     },
+     "execution_count": 27,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "X_train.dtypes"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "id": "altered-payday",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def get_patient_ids(df):\n",
+    "    return df.index.to_numpy().squeeze().tolist()\n",
+    "\n",
+    "scaler = MinMaxScaler()\n",
+    "#scaler = StandardScaler()\n",
+    "\n",
+    "ids_train = get_patient_ids(X_train)\n",
+    "X_train = scaler.fit_transform(X_train.to_numpy())\n",
+    "y_train = y_train.to_numpy().squeeze()\n",
+    "\n",
+    "ids_valid = get_patient_ids(X_valid)\n",
+    "X_valid = scaler.transform(X_valid.to_numpy())\n",
+    "y_valid = y_valid.to_numpy().squeeze()\n",
+    "\n",
+    "ids_test = get_patient_ids(X_test)\n",
+    "X_test = scaler.transform(X_test.to_numpy())\n",
+    "y_test = y_test.to_numpy().squeeze()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 29,
+   "id": "78b2ce16",
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(312, 105, 120)"
+      ]
+     },
+     "execution_count": 29,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(ids_train), len(ids_valid), len(ids_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 30,
+   "id": "stuffed-connecticut",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "(312, 11) (312,) 312\n",
+      "(105, 11) (105,) 105\n",
+      "(120, 11) (120,) 120\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(X_train.shape, y_train.shape, len(ids_train))\n",
+    "print(X_valid.shape, y_valid.shape, len(ids_valid))\n",
+    "print(X_test.shape, y_test.shape, len(ids_test))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 32,
+   "id": "instant-seating",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{'Age': 4.438642485192933, 'Sex': -0.5380532156403621, 'Temp_C': 0.9903044054257356, 'Cough': -0.09706348594397517, 'DifficultyInBreathing': 0.7928339789391249, 'WBC': 0.9729638366525475, 'CRP': 1.5618131479569666, 'LDH': 16.805483182146066, 'PaO2': -0.1568112199646922, 'CardiovascularDisease': -0.2591747799837606, 'RespiratoryFailure': -0.7772138167663587}\n",
+      "\n",
+      "significant coeffs {'Age': 4.438642485192933, 'Sex': -0.5380532156403621, 'Temp_C': 0.9903044054257356, 'Cough': -0.09706348594397517, 'DifficultyInBreathing': 0.7928339789391249, 'WBC': 0.9729638366525475, 'CRP': 1.5618131479569666, 'LDH': 16.805483182146066, 'PaO2': -0.1568112199646922, 'CardiovascularDisease': -0.2591747799837606, 'RespiratoryFailure': -0.7772138167663587}\n",
+      "\n",
+      "Training performance\n",
+      "=====\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.77      0.79      0.78       162\n",
+      "           1       0.77      0.74      0.75       150\n",
+      "\n",
+      "    accuracy                           0.77       312\n",
+      "   macro avg       0.77      0.77      0.77       312\n",
+      "weighted avg       0.77      0.77      0.77       312\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[128  34]\n",
+      " [ 39 111]]\n",
+      "\n",
+      "-AUC 0.7650617283950617\n",
+      "\n",
+      "Validation performance\n",
+      "=====\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.74      0.80      0.77        54\n",
+      "           1       0.77      0.71      0.73        51\n",
+      "\n",
+      "    accuracy                           0.75       105\n",
+      "   macro avg       0.75      0.75      0.75       105\n",
+      "weighted avg       0.75      0.75      0.75       105\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[43 11]\n",
+      " [15 36]]\n",
+      "\n",
+      "-AUC 0.7510893246187365\n",
+      "\n",
+      "Test performance\n",
+      "=====\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.39      0.89      0.54        35\n",
+      "           1       0.90      0.42      0.58        85\n",
+      "\n",
+      "    accuracy                           0.56       120\n",
+      "   macro avg       0.64      0.65      0.56       120\n",
+      "weighted avg       0.75      0.56      0.57       120\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[31  4]\n",
+      " [49 36]]\n",
+      "\n",
+      "-AUC 0.6546218487394958\n"
+     ]
+    }
+   ],
+   "source": [
+    "## simple logistic regression\n",
+    "from sklearn.linear_model import LogisticRegression\n",
+    "from sklearn.ensemble import RandomForestClassifier\n",
+    "from sklearn.metrics import classification_report, roc_auc_score, roc_curve, confusion_matrix\n",
+    "\n",
+    "#model = RandomForestClassifier() \n",
+    "model = LogisticRegression(penalty=\"none\", solver=\"lbfgs\")\n",
+    "\n",
+    "model.fit(X_train, y_train)\n",
+    "if hasattr(model, \"coef_\"):\n",
+    "    coeffs = dict(zip(feat_names, model.coef_[0]))\n",
+    "    print(coeffs)\n",
+    "    print(\"\\nsignificant coeffs\", {k:v for k,v in coeffs.items() if np.abs(v) > .01})\n",
+    "print()\n",
+    "\n",
+    "def evaluate_model(model, X, y):\n",
+    "    pred_y = model.predict(X)\n",
+    "\n",
+    "    print(\"-Classification metrics\\n\", classification_report(y, pred_y))\n",
+    "    print()\n",
+    "    print(\"-Confusion matrix\\n\", confusion_matrix(y, pred_y))\n",
+    "    print()\n",
+    "    print(\"-AUC\", roc_auc_score(y, pred_y))\n",
+    "    \n",
+    "\n",
+    "def make_prediction_df(model, X, patient_ids, method_name):\n",
+    "    cls_1_index = np.where(model.classes_ == 1)[0][0]\n",
+    "    pred_score = model.predict_proba(X)[:, cls_1_index]\n",
+    "    return pd.DataFrame({\"PatientID\": patient_ids, f\"prediction_{method_name}\": pred_score})\n",
+    "    \n",
+    "    \n",
+    "print(\"Training performance\\n=====\")\n",
+    "evaluate_model(model, X_train, y_train)\n",
+    "\n",
+    "print(\"\\nValidation performance\\n=====\")\n",
+    "evaluate_model(model, X_valid, y_valid)\n",
+    "\n",
+    "print(\"\\nTest performance\\n=====\")\n",
+    "evaluate_model(model, X_test, y_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "3b488c20",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model_name = \"Clinical-logistic-regression-nopenalty-lbfgs-trainedOn312Patients\"\n",
+    "pred_train = make_prediction_df(model, X_train, patient_ids=ids_train, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_valid = make_prediction_df(model, X_valid, patient_ids=ids_valid, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_test = make_prediction_df(model, X_test, patient_ids=ids_test, \n",
+    "                   method_name=model_name)\n",
+    "pred_test.to_csv(f\"/home/starke88/git/covid_data_challenge/clinical_model/predictions_{model_name}.csv\", index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a3c1d937",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "chemical-organ",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "if hasattr(model, \"feature_importances_\"):\n",
+    "    imp = dict(zip(feat_names, model.feature_importances_))\n",
+    "    for k, v in imp.items():\n",
+    "        print(k,v)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 33,
+   "id": "scientific-cancer",
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Fitting 5 folds for each of 4 candidates, totalling 20 fits\n",
+      ": grid-search best parameters {'fit__class_weight': 'balanced', 'fit__penalty': 'l1'}\n",
+      ": grid-search best score 0.7469022017409115\n",
+      "\n",
+      "\n",
+      "Training performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.76      0.78      0.77       162\n",
+      "           1       0.76      0.73      0.74       150\n",
+      "\n",
+      "    accuracy                           0.76       312\n",
+      "   macro avg       0.76      0.76      0.76       312\n",
+      "weighted avg       0.76      0.76      0.76       312\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[127  35]\n",
+      " [ 41 109]]\n",
+      "\n",
+      "-AUC 0.7553086419753087\n",
+      "\n",
+      "Validation performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.75      0.78      0.76        54\n",
+      "           1       0.76      0.73      0.74        51\n",
+      "\n",
+      "    accuracy                           0.75       105\n",
+      "   macro avg       0.75      0.75      0.75       105\n",
+      "weighted avg       0.75      0.75      0.75       105\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[42 12]\n",
+      " [14 37]]\n",
+      "\n",
+      "-AUC 0.7516339869281046\n",
+      "\n",
+      "Test performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.38      0.91      0.53        35\n",
+      "           1       0.91      0.38      0.53        85\n",
+      "\n",
+      "    accuracy                           0.53       120\n",
+      "   macro avg       0.65      0.65      0.53       120\n",
+      "weighted avg       0.76      0.53      0.53       120\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[32  3]\n",
+      " [53 32]]\n",
+      "\n",
+      "-AUC 0.6453781512605041\n"
+     ]
+    }
+   ],
+   "source": [
+    "from sklearn.pipeline import Pipeline\n",
+    "from sklearn.model_selection import GridSearchCV, KFold\n",
+    "\n",
+    "# l1 or l2\n",
+    "logistic_pipe = Pipeline([\n",
+    "    #(\"scaling\", MinMaxScaler()),\n",
+    "    (\"fit\", LogisticRegression(solver=\"liblinear\"))])\n",
+    "\n",
+    "logistic_grid = {\n",
+    "    #\"scaling\": [\"passthrough\", MinMaxScaler(), StandardScaler()],\n",
+    "    \"fit__penalty\": [\"l1\", \"l2\"],\n",
+    "    \"fit__class_weight\": [None, \"balanced\"]}\n",
+    "\n",
+    "\n",
+    "searcher = GridSearchCV(\n",
+    "            estimator=logistic_pipe,\n",
+    "            param_grid=logistic_grid,\n",
+    "            cv=KFold(n_splits=5, shuffle=False),\n",
+    "            #scoring=\"neg_mean_absolute_error\",\n",
+    "            verbose=1)\n",
+    "searcher.fit(X_train, y_train)\n",
+    "print(\": grid-search best parameters\", searcher.best_params_)\n",
+    "print(\": grid-search best score\", searcher.best_score_)\n",
+    "print()\n",
+    "\n",
+    "print(\"\\nTraining performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_train, y_train)\n",
+    "print(\"\\nValidation performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_valid, y_valid)\n",
+    "print(\"\\nTest performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_test, y_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "daf5b719",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model_name = \"Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn312Patients\"\n",
+    "model = searcher.best_estimator_\n",
+    "\n",
+    "pred_train = make_prediction_df(model, X_train, patient_ids=ids_train, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_valid = make_prediction_df(model, X_valid, patient_ids=ids_valid, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_test = make_prediction_df(model, X_test, patient_ids=ids_test, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_test.to_csv(f\"/home/starke88/git/covid_data_challenge/clinical_model/predictions_{model_name}.csv\", index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "id": "3c3165bc",
+   "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>PatientID</th>\n",
+       "      <th>prediction_Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn312Patients</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>P_102</td>\n",
+       "      <td>0.335200</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>P_117</td>\n",
+       "      <td>0.068277</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>P_16</td>\n",
+       "      <td>0.135848</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>P_118</td>\n",
+       "      <td>0.797524</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>P_114</td>\n",
+       "      <td>0.428291</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>115</th>\n",
+       "      <td>P_88</td>\n",
+       "      <td>0.204731</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>116</th>\n",
+       "      <td>P_92</td>\n",
+       "      <td>0.058076</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>117</th>\n",
+       "      <td>P_86</td>\n",
+       "      <td>0.494830</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>118</th>\n",
+       "      <td>P_9</td>\n",
+       "      <td>0.677815</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>119</th>\n",
+       "      <td>P_90</td>\n",
+       "      <td>0.321286</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>120 rows × 2 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "    PatientID  \\\n",
+       "0       P_102   \n",
+       "1       P_117   \n",
+       "2        P_16   \n",
+       "3       P_118   \n",
+       "4       P_114   \n",
+       "..        ...   \n",
+       "115      P_88   \n",
+       "116      P_92   \n",
+       "117      P_86   \n",
+       "118       P_9   \n",
+       "119      P_90   \n",
+       "\n",
+       "     prediction_Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn312Patients  \n",
+       "0                                             0.335200                                                    \n",
+       "1                                             0.068277                                                    \n",
+       "2                                             0.135848                                                    \n",
+       "3                                             0.797524                                                    \n",
+       "4                                             0.428291                                                    \n",
+       "..                                                 ...                                                    \n",
+       "115                                           0.204731                                                    \n",
+       "116                                           0.058076                                                    \n",
+       "117                                           0.494830                                                    \n",
+       "118                                           0.677815                                                    \n",
+       "119                                           0.321286                                                    \n",
+       "\n",
+       "[120 rows x 2 columns]"
+      ]
+     },
+     "execution_count": 24,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "pred_test"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 34,
+   "id": "dominican-colonial",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Fitting 5 folds for each of 400 candidates, totalling 2000 fits\n",
+      ": grid-search best parameters {'fit__max_depth': 5, 'fit__max_leaf_nodes': 20, 'fit__min_impurity_decrease': 0, 'fit__n_estimators': 10}\n",
+      ": grid-search best score 0.7596006144393241\n",
+      "\n",
+      "\n",
+      "Training performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.89      0.84      0.87       162\n",
+      "           1       0.84      0.89      0.86       150\n",
+      "\n",
+      "    accuracy                           0.87       312\n",
+      "   macro avg       0.87      0.87      0.87       312\n",
+      "weighted avg       0.87      0.87      0.87       312\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[136  26]\n",
+      " [ 16 134]]\n",
+      "\n",
+      "-AUC 0.8664197530864197\n",
+      "\n",
+      "Validation performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.81      0.72      0.76        54\n",
+      "           1       0.74      0.82      0.78        51\n",
+      "\n",
+      "    accuracy                           0.77       105\n",
+      "   macro avg       0.77      0.77      0.77       105\n",
+      "weighted avg       0.78      0.77      0.77       105\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[39 15]\n",
+      " [ 9 42]]\n",
+      "\n",
+      "-AUC 0.7728758169934641\n",
+      "\n",
+      "Test performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.41      0.89      0.56        35\n",
+      "           1       0.91      0.47      0.62        85\n",
+      "\n",
+      "    accuracy                           0.59       120\n",
+      "   macro avg       0.66      0.68      0.59       120\n",
+      "weighted avg       0.76      0.59      0.60       120\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[31  4]\n",
+      " [45 40]]\n",
+      "\n",
+      "-AUC 0.6781512605042017\n"
+     ]
+    }
+   ],
+   "source": [
+    "rf_pipe = Pipeline([\n",
+    "    #(\"scaling\", MinMaxScaler()),\n",
+    "    (\"fit\", RandomForestClassifier(n_estimators=100,\n",
+    "                                  # max_depth=5,\n",
+    "                                  # max_leaf_nodes=20,\n",
+    "                                  # min_impurity_decrease=.1\n",
+    "                                  random_state=42,\n",
+    "                                  ))\n",
+    "])\n",
+    "rf_grid = {\n",
+    "    #\"scaling\": [MinMaxScaler(), StandardScaler()],\n",
+    "    \"fit__n_estimators\": [5, 10, 20, 100, 200],\n",
+    "    \"fit__max_depth\": [None, 5, 10, 20, 50],\n",
+    "    \"fit__max_leaf_nodes\": [None, 20, 50, 100],\n",
+    "    \"fit__min_impurity_decrease\": [0, 0.1, .5, 1]    }\n",
+    "\n",
+    "searcher = GridSearchCV(\n",
+    "            estimator=rf_pipe,\n",
+    "            param_grid=rf_grid,\n",
+    "            cv=KFold(n_splits=5, shuffle=False),\n",
+    "            #scoring=\"neg_mean_absolute_error\",\n",
+    "            verbose=1)\n",
+    "searcher.fit(X_train, y_train)\n",
+    "\n",
+    "print(\": grid-search best parameters\", searcher.best_params_)\n",
+    "print(\": grid-search best score\", searcher.best_score_)\n",
+    "print()\n",
+    "print(\"\\nTraining performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_train, y_train)\n",
+    "print(\"\\nValidation performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_valid, y_valid)\n",
+    "print(\"\\nTest performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_test, y_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 26,
+   "id": "435a051d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model_name = \"Clinical-randomforest-10estimators-trainedOn312Patients\"\n",
+    "model = searcher.best_estimator_\n",
+    "\n",
+    "pred_train = make_prediction_df(model, X_train, patient_ids=ids_train, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_valid = make_prediction_df(model, X_valid, patient_ids=ids_valid, \n",
+    "                   method_name=model_name)\n",
+    "pred_test = make_prediction_df(model, X_test, patient_ids=ids_test, \n",
+    "                   method_name=model_name)\n",
+    "pred_test.to_csv(f\"/home/starke88/git/covid_data_challenge/clinical_model/predictions_{model_name}.csv\", index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "aca20c17",
+   "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": "3.8.8"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/clinical_model/clinical_data_with_imputed_values.ipynb b/clinical_model/clinical_data_with_imputed_values.ipynb
new file mode 100644
index 0000000..e531bd0
--- /dev/null
+++ b/clinical_model/clinical_data_with_imputed_values.ipynb
@@ -0,0 +1,2319 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "electoral-korean",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "import pathlib\n",
+    "import numpy as np"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "9aedf0e5",
+   "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>Age</th>\n",
+       "      <th>Sex</th>\n",
+       "      <th>Temp_C</th>\n",
+       "      <th>Cough</th>\n",
+       "      <th>DifficultyInBreathing</th>\n",
+       "      <th>WBC</th>\n",
+       "      <th>CRP</th>\n",
+       "      <th>LDH</th>\n",
+       "      <th>Ox_percentage</th>\n",
+       "      <th>PaO2</th>\n",
+       "      <th>pH</th>\n",
+       "      <th>CardiovascularDisease</th>\n",
+       "      <th>RespiratoryFailure</th>\n",
+       "      <th>Prognosis</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>PatientID</th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>P_131</th>\n",
+       "      <td>35.913889</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>39.300000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.76</td>\n",
+       "      <td>43.40</td>\n",
+       "      <td>387.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>78.359259</td>\n",
+       "      <td>7.462222</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>MILD</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_132</th>\n",
+       "      <td>57.266667</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>11.48</td>\n",
+       "      <td>64.00</td>\n",
+       "      <td>338.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>75.000000</td>\n",
+       "      <td>7.420000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>MILD</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_195</th>\n",
+       "      <td>79.263889</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.800000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.21</td>\n",
+       "      <td>115.30</td>\n",
+       "      <td>356.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>63.000000</td>\n",
+       "      <td>7.390000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>SEVERE</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_193</th>\n",
+       "      <td>82.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>7.28</td>\n",
+       "      <td>149.30</td>\n",
+       "      <td>482.000000</td>\n",
+       "      <td>97.000000</td>\n",
+       "      <td>68.000000</td>\n",
+       "      <td>7.460000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>SEVERE</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_140</th>\n",
+       "      <td>60.791667</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>37.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.37</td>\n",
+       "      <td>20.70</td>\n",
+       "      <td>466.481481</td>\n",
+       "      <td>93.000000</td>\n",
+       "      <td>66.962963</td>\n",
+       "      <td>7.452963</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>MILD</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_12</th>\n",
+       "      <td>51.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.533333</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>14.30</td>\n",
+       "      <td>22.79</td>\n",
+       "      <td>368.000000</td>\n",
+       "      <td>93.703704</td>\n",
+       "      <td>69.266667</td>\n",
+       "      <td>7.452963</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>SEVERE</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_8</th>\n",
+       "      <td>57.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.274074</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.10</td>\n",
+       "      <td>8.93</td>\n",
+       "      <td>451.000000</td>\n",
+       "      <td>93.551852</td>\n",
+       "      <td>70.651852</td>\n",
+       "      <td>7.451852</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>SEVERE</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_10</th>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.659259</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>7.30</td>\n",
+       "      <td>0.23</td>\n",
+       "      <td>212.000000</td>\n",
+       "      <td>96.814815</td>\n",
+       "      <td>83.962963</td>\n",
+       "      <td>7.430370</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>MILD</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_26</th>\n",
+       "      <td>92.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>13.50</td>\n",
+       "      <td>3.77</td>\n",
+       "      <td>358.000000</td>\n",
+       "      <td>92.703704</td>\n",
+       "      <td>75.700000</td>\n",
+       "      <td>7.360000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>SEVERE</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_146</th>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>38.400000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.70</td>\n",
+       "      <td>0.02</td>\n",
+       "      <td>149.000000</td>\n",
+       "      <td>94.551852</td>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>7.460000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>MILD</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>863 rows × 14 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                 Age  Sex     Temp_C  Cough  DifficultyInBreathing    WBC  \\\n",
+       "PatientID                                                                   \n",
+       "P_131      35.913889  0.0  39.300000    0.0                    0.0   5.76   \n",
+       "P_132      57.266667  0.0  37.000000    1.0                    0.0  11.48   \n",
+       "P_195      79.263889  0.0  37.800000    0.0                    0.0   6.21   \n",
+       "P_193      82.000000  0.0  38.000000    0.0                    0.0   7.28   \n",
+       "P_140      60.791667  1.0  37.000000    0.0                    0.0   6.37   \n",
+       "...              ...  ...        ...    ...                    ...    ...   \n",
+       "P_1_12     51.000000  0.0  37.533333    1.0                    1.0  14.30   \n",
+       "P_1_8      57.000000  0.0  37.274074    0.0                    0.0   5.10   \n",
+       "P_1_10     38.000000  0.0  37.659259    1.0                    1.0   7.30   \n",
+       "P_1_26     92.000000  1.0  38.000000    1.0                    1.0  13.50   \n",
+       "P_1_146    80.000000  0.0  38.400000    1.0                    0.0   6.70   \n",
+       "\n",
+       "              CRP         LDH  Ox_percentage       PaO2        pH  \\\n",
+       "PatientID                                                           \n",
+       "P_131       43.40  387.000000      94.000000  78.359259  7.462222   \n",
+       "P_132       64.00  338.000000      94.000000  75.000000  7.420000   \n",
+       "P_195      115.30  356.000000      94.000000  63.000000  7.390000   \n",
+       "P_193      149.30  482.000000      97.000000  68.000000  7.460000   \n",
+       "P_140       20.70  466.481481      93.000000  66.962963  7.452963   \n",
+       "...           ...         ...            ...        ...       ...   \n",
+       "P_1_12      22.79  368.000000      93.703704  69.266667  7.452963   \n",
+       "P_1_8        8.93  451.000000      93.551852  70.651852  7.451852   \n",
+       "P_1_10       0.23  212.000000      96.814815  83.962963  7.430370   \n",
+       "P_1_26       3.77  358.000000      92.703704  75.700000  7.360000   \n",
+       "P_1_146      0.02  149.000000      94.551852  80.000000  7.460000   \n",
+       "\n",
+       "           CardiovascularDisease  RespiratoryFailure Prognosis  \n",
+       "PatientID                                                       \n",
+       "P_131                        0.0                -1.0      MILD  \n",
+       "P_132                        0.0                -1.0      MILD  \n",
+       "P_195                        1.0                -1.0    SEVERE  \n",
+       "P_193                        0.0                -1.0    SEVERE  \n",
+       "P_140                        0.0                -1.0      MILD  \n",
+       "...                          ...                 ...       ...  \n",
+       "P_1_12                       0.0                 0.0    SEVERE  \n",
+       "P_1_8                        0.0                 0.0    SEVERE  \n",
+       "P_1_10                       0.0                 0.0      MILD  \n",
+       "P_1_26                       0.0                 0.0    SEVERE  \n",
+       "P_1_146                      1.0                 0.0      MILD  \n",
+       "\n",
+       "[863 rows x 14 columns]"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "train_path = pathlib.Path(\"/home/starke88/data/haicu/covid_data_challenge_2021/trainSet_filled2.txt\")\n",
+    "df_train = pd.read_csv(train_path)\n",
+    "df_train.set_index(\"PatientID\", inplace=True)\n",
+    "#df_train.drop([\"Unnamed: 0\"], axis=1, inplace=True)\n",
+    "\n",
+    "#df_old = pd.read_csv(pathlib.Path(\"/home/starke88/data/haicu/covid_data_challenge_2021/trainSet.txt\"))\n",
+    "#df_old.set_index(\"PatientID\", inplace=True)\n",
+    "#df_old = df_old[[\"Prognosis\"]]\n",
+    "\n",
+    "#df_train = pd.concat([df_train, df_old], axis=1)\n",
+    "\n",
+    "df_train"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "fc4bbb3d",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "boolean-bacteria",
+   "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>Age</th>\n",
+       "      <th>Sex</th>\n",
+       "      <th>Temp_C</th>\n",
+       "      <th>Cough</th>\n",
+       "      <th>DifficultyInBreathing</th>\n",
+       "      <th>WBC</th>\n",
+       "      <th>CRP</th>\n",
+       "      <th>LDH</th>\n",
+       "      <th>Ox_percentage</th>\n",
+       "      <th>PaO2</th>\n",
+       "      <th>pH</th>\n",
+       "      <th>CardiovascularDisease</th>\n",
+       "      <th>RespiratoryFailure</th>\n",
+       "      <th>Prognosis</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>PatientID</th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>P_131</th>\n",
+       "      <td>35.913889</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>39.300000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.76</td>\n",
+       "      <td>43.40</td>\n",
+       "      <td>387.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>78.359259</td>\n",
+       "      <td>7.462222</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_132</th>\n",
+       "      <td>57.266667</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>11.48</td>\n",
+       "      <td>64.00</td>\n",
+       "      <td>338.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>75.000000</td>\n",
+       "      <td>7.420000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_195</th>\n",
+       "      <td>79.263889</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.800000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.21</td>\n",
+       "      <td>115.30</td>\n",
+       "      <td>356.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>63.000000</td>\n",
+       "      <td>7.390000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_193</th>\n",
+       "      <td>82.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>7.28</td>\n",
+       "      <td>149.30</td>\n",
+       "      <td>482.000000</td>\n",
+       "      <td>97.000000</td>\n",
+       "      <td>68.000000</td>\n",
+       "      <td>7.460000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_140</th>\n",
+       "      <td>60.791667</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>37.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.37</td>\n",
+       "      <td>20.70</td>\n",
+       "      <td>466.481481</td>\n",
+       "      <td>93.000000</td>\n",
+       "      <td>66.962963</td>\n",
+       "      <td>7.452963</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_12</th>\n",
+       "      <td>51.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.533333</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>14.30</td>\n",
+       "      <td>22.79</td>\n",
+       "      <td>368.000000</td>\n",
+       "      <td>93.703704</td>\n",
+       "      <td>69.266667</td>\n",
+       "      <td>7.452963</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_8</th>\n",
+       "      <td>57.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.274074</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.10</td>\n",
+       "      <td>8.93</td>\n",
+       "      <td>451.000000</td>\n",
+       "      <td>93.551852</td>\n",
+       "      <td>70.651852</td>\n",
+       "      <td>7.451852</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_10</th>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.659259</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>7.30</td>\n",
+       "      <td>0.23</td>\n",
+       "      <td>212.000000</td>\n",
+       "      <td>96.814815</td>\n",
+       "      <td>83.962963</td>\n",
+       "      <td>7.430370</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_26</th>\n",
+       "      <td>92.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>13.50</td>\n",
+       "      <td>3.77</td>\n",
+       "      <td>358.000000</td>\n",
+       "      <td>92.703704</td>\n",
+       "      <td>75.700000</td>\n",
+       "      <td>7.360000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_146</th>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>38.400000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.70</td>\n",
+       "      <td>0.02</td>\n",
+       "      <td>149.000000</td>\n",
+       "      <td>94.551852</td>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>7.460000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>863 rows × 14 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                 Age  Sex     Temp_C  Cough  DifficultyInBreathing    WBC  \\\n",
+       "PatientID                                                                   \n",
+       "P_131      35.913889  0.0  39.300000    0.0                    0.0   5.76   \n",
+       "P_132      57.266667  0.0  37.000000    1.0                    0.0  11.48   \n",
+       "P_195      79.263889  0.0  37.800000    0.0                    0.0   6.21   \n",
+       "P_193      82.000000  0.0  38.000000    0.0                    0.0   7.28   \n",
+       "P_140      60.791667  1.0  37.000000    0.0                    0.0   6.37   \n",
+       "...              ...  ...        ...    ...                    ...    ...   \n",
+       "P_1_12     51.000000  0.0  37.533333    1.0                    1.0  14.30   \n",
+       "P_1_8      57.000000  0.0  37.274074    0.0                    0.0   5.10   \n",
+       "P_1_10     38.000000  0.0  37.659259    1.0                    1.0   7.30   \n",
+       "P_1_26     92.000000  1.0  38.000000    1.0                    1.0  13.50   \n",
+       "P_1_146    80.000000  0.0  38.400000    1.0                    0.0   6.70   \n",
+       "\n",
+       "              CRP         LDH  Ox_percentage       PaO2        pH  \\\n",
+       "PatientID                                                           \n",
+       "P_131       43.40  387.000000      94.000000  78.359259  7.462222   \n",
+       "P_132       64.00  338.000000      94.000000  75.000000  7.420000   \n",
+       "P_195      115.30  356.000000      94.000000  63.000000  7.390000   \n",
+       "P_193      149.30  482.000000      97.000000  68.000000  7.460000   \n",
+       "P_140       20.70  466.481481      93.000000  66.962963  7.452963   \n",
+       "...           ...         ...            ...        ...       ...   \n",
+       "P_1_12      22.79  368.000000      93.703704  69.266667  7.452963   \n",
+       "P_1_8        8.93  451.000000      93.551852  70.651852  7.451852   \n",
+       "P_1_10       0.23  212.000000      96.814815  83.962963  7.430370   \n",
+       "P_1_26       3.77  358.000000      92.703704  75.700000  7.360000   \n",
+       "P_1_146      0.02  149.000000      94.551852  80.000000  7.460000   \n",
+       "\n",
+       "           CardiovascularDisease  RespiratoryFailure  Prognosis  \n",
+       "PatientID                                                        \n",
+       "P_131                        0.0                -1.0          0  \n",
+       "P_132                        0.0                -1.0          0  \n",
+       "P_195                        1.0                -1.0          1  \n",
+       "P_193                        0.0                -1.0          1  \n",
+       "P_140                        0.0                -1.0          0  \n",
+       "...                          ...                 ...        ...  \n",
+       "P_1_12                       0.0                 0.0          1  \n",
+       "P_1_8                        0.0                 0.0          1  \n",
+       "P_1_10                       0.0                 0.0          0  \n",
+       "P_1_26                       0.0                 0.0          1  \n",
+       "P_1_146                      1.0                 0.0          0  \n",
+       "\n",
+       "[863 rows x 14 columns]"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "\n",
+    "#df_train.drop([\"ImageFile\"], axis=1, inplace=True)\n",
+    "df_train[\"Prognosis\"] = (df_train[\"Prognosis\"] == \"SEVERE\").astype(np.uint8)\n",
+    "\n",
+    "# one-hot encode the hospital\n",
+    "#foo = pd.get_dummies(df_train[\"Hospital\"], prefix=\"Hospital\")\n",
+    "#df_train = pd.concat([df_train, foo], axis=1)\n",
+    "#df_train = df_train.drop([\"Hospital\"], axis=1)\n",
+    "\n",
+    "\n",
+    "df_train"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "fantastic-struggle",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "(863, 14)"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# this is how many complete samples we have\n",
+    "df_train.dropna().shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "18852aa5",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "ongoing-dance",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Age \t\t\t 0 / 863\n",
+      "Sex \t\t\t 0 / 863\n",
+      "Temp_C \t\t\t 0 / 863\n",
+      "Cough \t\t\t 0 / 863\n",
+      "DifficultyInBreathing \t\t\t 0 / 863\n",
+      "WBC \t\t\t 0 / 863\n",
+      "CRP \t\t\t 0 / 863\n",
+      "LDH \t\t\t 0 / 863\n",
+      "Ox_percentage \t\t\t 0 / 863\n",
+      "PaO2 \t\t\t 0 / 863\n",
+      "pH \t\t\t 0 / 863\n",
+      "CardiovascularDisease \t\t\t 0 / 863\n",
+      "RespiratoryFailure \t\t\t 0 / 863\n",
+      "Prognosis \t\t\t 0 / 863\n",
+      "\n",
+      "Going to discard features with more than 200 nans: []\n"
+     ]
+    }
+   ],
+   "source": [
+    "# how many samples are missing for individual columns\n",
+    "discard = []\n",
+    "discard_thresh = 200\n",
+    "for c in df_train.columns:\n",
+    "    n_missing = df_train[c].isna().sum()\n",
+    "    print(c, \"\\t\\t\\t\", n_missing, \"/\", len(df_train))\n",
+    "    if n_missing > discard_thresh:\n",
+    "        discard.append(c)\n",
+    "        \n",
+    "print(f\"\\nGoing to discard features with more than {discard_thresh} nans:\", discard)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "induced-treat",
+   "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>Age</th>\n",
+       "      <th>Sex</th>\n",
+       "      <th>Temp_C</th>\n",
+       "      <th>Cough</th>\n",
+       "      <th>DifficultyInBreathing</th>\n",
+       "      <th>WBC</th>\n",
+       "      <th>CRP</th>\n",
+       "      <th>LDH</th>\n",
+       "      <th>Ox_percentage</th>\n",
+       "      <th>PaO2</th>\n",
+       "      <th>pH</th>\n",
+       "      <th>CardiovascularDisease</th>\n",
+       "      <th>RespiratoryFailure</th>\n",
+       "      <th>Prognosis</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>PatientID</th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>P_131</th>\n",
+       "      <td>35.913889</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>39.300000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.76</td>\n",
+       "      <td>43.40</td>\n",
+       "      <td>387.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>78.359259</td>\n",
+       "      <td>7.462222</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_132</th>\n",
+       "      <td>57.266667</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>11.48</td>\n",
+       "      <td>64.00</td>\n",
+       "      <td>338.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>75.000000</td>\n",
+       "      <td>7.420000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_195</th>\n",
+       "      <td>79.263889</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.800000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.21</td>\n",
+       "      <td>115.30</td>\n",
+       "      <td>356.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>63.000000</td>\n",
+       "      <td>7.390000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_193</th>\n",
+       "      <td>82.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>7.28</td>\n",
+       "      <td>149.30</td>\n",
+       "      <td>482.000000</td>\n",
+       "      <td>97.000000</td>\n",
+       "      <td>68.000000</td>\n",
+       "      <td>7.460000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_140</th>\n",
+       "      <td>60.791667</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>37.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.37</td>\n",
+       "      <td>20.70</td>\n",
+       "      <td>466.481481</td>\n",
+       "      <td>93.000000</td>\n",
+       "      <td>66.962963</td>\n",
+       "      <td>7.452963</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_12</th>\n",
+       "      <td>51.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.533333</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>14.30</td>\n",
+       "      <td>22.79</td>\n",
+       "      <td>368.000000</td>\n",
+       "      <td>93.703704</td>\n",
+       "      <td>69.266667</td>\n",
+       "      <td>7.452963</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_8</th>\n",
+       "      <td>57.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.274074</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.10</td>\n",
+       "      <td>8.93</td>\n",
+       "      <td>451.000000</td>\n",
+       "      <td>93.551852</td>\n",
+       "      <td>70.651852</td>\n",
+       "      <td>7.451852</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_10</th>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.659259</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>7.30</td>\n",
+       "      <td>0.23</td>\n",
+       "      <td>212.000000</td>\n",
+       "      <td>96.814815</td>\n",
+       "      <td>83.962963</td>\n",
+       "      <td>7.430370</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_26</th>\n",
+       "      <td>92.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>13.50</td>\n",
+       "      <td>3.77</td>\n",
+       "      <td>358.000000</td>\n",
+       "      <td>92.703704</td>\n",
+       "      <td>75.700000</td>\n",
+       "      <td>7.360000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_146</th>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>38.400000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.70</td>\n",
+       "      <td>0.02</td>\n",
+       "      <td>149.000000</td>\n",
+       "      <td>94.551852</td>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>7.460000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>863 rows × 14 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                 Age  Sex     Temp_C  Cough  DifficultyInBreathing    WBC  \\\n",
+       "PatientID                                                                   \n",
+       "P_131      35.913889  0.0  39.300000    0.0                    0.0   5.76   \n",
+       "P_132      57.266667  0.0  37.000000    1.0                    0.0  11.48   \n",
+       "P_195      79.263889  0.0  37.800000    0.0                    0.0   6.21   \n",
+       "P_193      82.000000  0.0  38.000000    0.0                    0.0   7.28   \n",
+       "P_140      60.791667  1.0  37.000000    0.0                    0.0   6.37   \n",
+       "...              ...  ...        ...    ...                    ...    ...   \n",
+       "P_1_12     51.000000  0.0  37.533333    1.0                    1.0  14.30   \n",
+       "P_1_8      57.000000  0.0  37.274074    0.0                    0.0   5.10   \n",
+       "P_1_10     38.000000  0.0  37.659259    1.0                    1.0   7.30   \n",
+       "P_1_26     92.000000  1.0  38.000000    1.0                    1.0  13.50   \n",
+       "P_1_146    80.000000  0.0  38.400000    1.0                    0.0   6.70   \n",
+       "\n",
+       "              CRP         LDH  Ox_percentage       PaO2        pH  \\\n",
+       "PatientID                                                           \n",
+       "P_131       43.40  387.000000      94.000000  78.359259  7.462222   \n",
+       "P_132       64.00  338.000000      94.000000  75.000000  7.420000   \n",
+       "P_195      115.30  356.000000      94.000000  63.000000  7.390000   \n",
+       "P_193      149.30  482.000000      97.000000  68.000000  7.460000   \n",
+       "P_140       20.70  466.481481      93.000000  66.962963  7.452963   \n",
+       "...           ...         ...            ...        ...       ...   \n",
+       "P_1_12      22.79  368.000000      93.703704  69.266667  7.452963   \n",
+       "P_1_8        8.93  451.000000      93.551852  70.651852  7.451852   \n",
+       "P_1_10       0.23  212.000000      96.814815  83.962963  7.430370   \n",
+       "P_1_26       3.77  358.000000      92.703704  75.700000  7.360000   \n",
+       "P_1_146      0.02  149.000000      94.551852  80.000000  7.460000   \n",
+       "\n",
+       "           CardiovascularDisease  RespiratoryFailure  Prognosis  \n",
+       "PatientID                                                        \n",
+       "P_131                        0.0                -1.0          0  \n",
+       "P_132                        0.0                -1.0          0  \n",
+       "P_195                        1.0                -1.0          1  \n",
+       "P_193                        0.0                -1.0          1  \n",
+       "P_140                        0.0                -1.0          0  \n",
+       "...                          ...                 ...        ...  \n",
+       "P_1_12                       0.0                 0.0          1  \n",
+       "P_1_8                        0.0                 0.0          1  \n",
+       "P_1_10                       0.0                 0.0          0  \n",
+       "P_1_26                       0.0                 0.0          1  \n",
+       "P_1_146                      1.0                 0.0          0  \n",
+       "\n",
+       "[863 rows x 14 columns]"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "# since all values are imputed we should not have anything missing here anymore\n",
+    "df_reduced = df_train.drop(discard, axis=1)\n",
+    "df_reduced"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "bizarre-college",
+   "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>Age</th>\n",
+       "      <th>Sex</th>\n",
+       "      <th>Temp_C</th>\n",
+       "      <th>Cough</th>\n",
+       "      <th>DifficultyInBreathing</th>\n",
+       "      <th>WBC</th>\n",
+       "      <th>CRP</th>\n",
+       "      <th>LDH</th>\n",
+       "      <th>Ox_percentage</th>\n",
+       "      <th>PaO2</th>\n",
+       "      <th>pH</th>\n",
+       "      <th>CardiovascularDisease</th>\n",
+       "      <th>RespiratoryFailure</th>\n",
+       "      <th>Prognosis</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>count</th>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "      <td>863.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>mean</th>\n",
+       "      <td>64.443784</td>\n",
+       "      <td>0.337196</td>\n",
+       "      <td>37.591352</td>\n",
+       "      <td>0.483198</td>\n",
+       "      <td>0.490151</td>\n",
+       "      <td>7.057532</td>\n",
+       "      <td>40.568481</td>\n",
+       "      <td>362.471459</td>\n",
+       "      <td>92.673928</td>\n",
+       "      <td>72.734359</td>\n",
+       "      <td>7.454625</td>\n",
+       "      <td>0.249131</td>\n",
+       "      <td>-0.171495</td>\n",
+       "      <td>0.492468</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>std</th>\n",
+       "      <td>15.044203</td>\n",
+       "      <td>0.473027</td>\n",
+       "      <td>0.884709</td>\n",
+       "      <td>0.511477</td>\n",
+       "      <td>0.509386</td>\n",
+       "      <td>3.514950</td>\n",
+       "      <td>65.736724</td>\n",
+       "      <td>217.412325</td>\n",
+       "      <td>6.040142</td>\n",
+       "      <td>23.537867</td>\n",
+       "      <td>0.049681</td>\n",
+       "      <td>0.481004</td>\n",
+       "      <td>0.409599</td>\n",
+       "      <td>0.500233</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>min</th>\n",
+       "      <td>18.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>35.400000</td>\n",
+       "      <td>-1.000000</td>\n",
+       "      <td>-1.000000</td>\n",
+       "      <td>0.200000</td>\n",
+       "      <td>0.010000</td>\n",
+       "      <td>0.020000</td>\n",
+       "      <td>50.000000</td>\n",
+       "      <td>23.000000</td>\n",
+       "      <td>7.048000</td>\n",
+       "      <td>-1.000000</td>\n",
+       "      <td>-1.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>25%</th>\n",
+       "      <td>54.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>37.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>4.660000</td>\n",
+       "      <td>5.970000</td>\n",
+       "      <td>251.000000</td>\n",
+       "      <td>91.153704</td>\n",
+       "      <td>61.900000</td>\n",
+       "      <td>7.437037</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>50%</th>\n",
+       "      <td>65.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>37.603704</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>6.270000</td>\n",
+       "      <td>15.420000</td>\n",
+       "      <td>321.370370</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>71.522222</td>\n",
+       "      <td>7.454741</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>75%</th>\n",
+       "      <td>77.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>8.400000</td>\n",
+       "      <td>42.700000</td>\n",
+       "      <td>405.500000</td>\n",
+       "      <td>96.000000</td>\n",
+       "      <td>78.990741</td>\n",
+       "      <td>7.480000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>0.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>max</th>\n",
+       "      <td>97.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>40.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>27.460000</td>\n",
+       "      <td>570.500000</td>\n",
+       "      <td>2903.000000</td>\n",
+       "      <td>100.000000</td>\n",
+       "      <td>285.000000</td>\n",
+       "      <td>7.990000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "      <td>1.000000</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "              Age         Sex      Temp_C       Cough  DifficultyInBreathing  \\\n",
+       "count  863.000000  863.000000  863.000000  863.000000             863.000000   \n",
+       "mean    64.443784    0.337196   37.591352    0.483198               0.490151   \n",
+       "std     15.044203    0.473027    0.884709    0.511477               0.509386   \n",
+       "min     18.000000    0.000000   35.400000   -1.000000              -1.000000   \n",
+       "25%     54.000000    0.000000   37.000000    0.000000               0.000000   \n",
+       "50%     65.000000    0.000000   37.603704    0.000000               0.000000   \n",
+       "75%     77.000000    1.000000   38.000000    1.000000               1.000000   \n",
+       "max     97.000000    1.000000   40.000000    1.000000               1.000000   \n",
+       "\n",
+       "              WBC         CRP          LDH  Ox_percentage        PaO2  \\\n",
+       "count  863.000000  863.000000   863.000000     863.000000  863.000000   \n",
+       "mean     7.057532   40.568481   362.471459      92.673928   72.734359   \n",
+       "std      3.514950   65.736724   217.412325       6.040142   23.537867   \n",
+       "min      0.200000    0.010000     0.020000      50.000000   23.000000   \n",
+       "25%      4.660000    5.970000   251.000000      91.153704   61.900000   \n",
+       "50%      6.270000   15.420000   321.370370      94.000000   71.522222   \n",
+       "75%      8.400000   42.700000   405.500000      96.000000   78.990741   \n",
+       "max     27.460000  570.500000  2903.000000     100.000000  285.000000   \n",
+       "\n",
+       "               pH  CardiovascularDisease  RespiratoryFailure   Prognosis  \n",
+       "count  863.000000             863.000000          863.000000  863.000000  \n",
+       "mean     7.454625               0.249131           -0.171495    0.492468  \n",
+       "std      0.049681               0.481004            0.409599    0.500233  \n",
+       "min      7.048000              -1.000000           -1.000000    0.000000  \n",
+       "25%      7.437037               0.000000            0.000000    0.000000  \n",
+       "50%      7.454741               0.000000            0.000000    0.000000  \n",
+       "75%      7.480000               1.000000            0.000000    1.000000  \n",
+       "max      7.990000               1.000000            1.000000    1.000000  "
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_reduced = df_reduced.dropna()\n",
+    "df_reduced.describe()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "compact-lawsuit",
+   "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>Age</th>\n",
+       "      <th>Sex</th>\n",
+       "      <th>Temp_C</th>\n",
+       "      <th>Cough</th>\n",
+       "      <th>DifficultyInBreathing</th>\n",
+       "      <th>WBC</th>\n",
+       "      <th>CRP</th>\n",
+       "      <th>LDH</th>\n",
+       "      <th>Ox_percentage</th>\n",
+       "      <th>PaO2</th>\n",
+       "      <th>pH</th>\n",
+       "      <th>CardiovascularDisease</th>\n",
+       "      <th>RespiratoryFailure</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>PatientID</th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>P_131</th>\n",
+       "      <td>35.913889</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>39.300000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.76</td>\n",
+       "      <td>43.40</td>\n",
+       "      <td>387.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>78.359259</td>\n",
+       "      <td>7.462222</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_132</th>\n",
+       "      <td>57.266667</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>11.48</td>\n",
+       "      <td>64.00</td>\n",
+       "      <td>338.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>75.000000</td>\n",
+       "      <td>7.420000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_195</th>\n",
+       "      <td>79.263889</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.800000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.21</td>\n",
+       "      <td>115.30</td>\n",
+       "      <td>356.000000</td>\n",
+       "      <td>94.000000</td>\n",
+       "      <td>63.000000</td>\n",
+       "      <td>7.390000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_193</th>\n",
+       "      <td>82.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>7.28</td>\n",
+       "      <td>149.30</td>\n",
+       "      <td>482.000000</td>\n",
+       "      <td>97.000000</td>\n",
+       "      <td>68.000000</td>\n",
+       "      <td>7.460000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_140</th>\n",
+       "      <td>60.791667</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>37.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.37</td>\n",
+       "      <td>20.70</td>\n",
+       "      <td>466.481481</td>\n",
+       "      <td>93.000000</td>\n",
+       "      <td>66.962963</td>\n",
+       "      <td>7.452963</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_12</th>\n",
+       "      <td>51.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.533333</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>14.30</td>\n",
+       "      <td>22.79</td>\n",
+       "      <td>368.000000</td>\n",
+       "      <td>93.703704</td>\n",
+       "      <td>69.266667</td>\n",
+       "      <td>7.452963</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_8</th>\n",
+       "      <td>57.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.274074</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>5.10</td>\n",
+       "      <td>8.93</td>\n",
+       "      <td>451.000000</td>\n",
+       "      <td>93.551852</td>\n",
+       "      <td>70.651852</td>\n",
+       "      <td>7.451852</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_10</th>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>37.659259</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>7.30</td>\n",
+       "      <td>0.23</td>\n",
+       "      <td>212.000000</td>\n",
+       "      <td>96.814815</td>\n",
+       "      <td>83.962963</td>\n",
+       "      <td>7.430370</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_26</th>\n",
+       "      <td>92.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>38.000000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>13.50</td>\n",
+       "      <td>3.77</td>\n",
+       "      <td>358.000000</td>\n",
+       "      <td>92.703704</td>\n",
+       "      <td>75.700000</td>\n",
+       "      <td>7.360000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_1_146</th>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>38.400000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "      <td>6.70</td>\n",
+       "      <td>0.02</td>\n",
+       "      <td>149.000000</td>\n",
+       "      <td>94.551852</td>\n",
+       "      <td>80.000000</td>\n",
+       "      <td>7.460000</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>863 rows × 13 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                 Age  Sex     Temp_C  Cough  DifficultyInBreathing    WBC  \\\n",
+       "PatientID                                                                   \n",
+       "P_131      35.913889  0.0  39.300000    0.0                    0.0   5.76   \n",
+       "P_132      57.266667  0.0  37.000000    1.0                    0.0  11.48   \n",
+       "P_195      79.263889  0.0  37.800000    0.0                    0.0   6.21   \n",
+       "P_193      82.000000  0.0  38.000000    0.0                    0.0   7.28   \n",
+       "P_140      60.791667  1.0  37.000000    0.0                    0.0   6.37   \n",
+       "...              ...  ...        ...    ...                    ...    ...   \n",
+       "P_1_12     51.000000  0.0  37.533333    1.0                    1.0  14.30   \n",
+       "P_1_8      57.000000  0.0  37.274074    0.0                    0.0   5.10   \n",
+       "P_1_10     38.000000  0.0  37.659259    1.0                    1.0   7.30   \n",
+       "P_1_26     92.000000  1.0  38.000000    1.0                    1.0  13.50   \n",
+       "P_1_146    80.000000  0.0  38.400000    1.0                    0.0   6.70   \n",
+       "\n",
+       "              CRP         LDH  Ox_percentage       PaO2        pH  \\\n",
+       "PatientID                                                           \n",
+       "P_131       43.40  387.000000      94.000000  78.359259  7.462222   \n",
+       "P_132       64.00  338.000000      94.000000  75.000000  7.420000   \n",
+       "P_195      115.30  356.000000      94.000000  63.000000  7.390000   \n",
+       "P_193      149.30  482.000000      97.000000  68.000000  7.460000   \n",
+       "P_140       20.70  466.481481      93.000000  66.962963  7.452963   \n",
+       "...           ...         ...            ...        ...       ...   \n",
+       "P_1_12      22.79  368.000000      93.703704  69.266667  7.452963   \n",
+       "P_1_8        8.93  451.000000      93.551852  70.651852  7.451852   \n",
+       "P_1_10       0.23  212.000000      96.814815  83.962963  7.430370   \n",
+       "P_1_26       3.77  358.000000      92.703704  75.700000  7.360000   \n",
+       "P_1_146      0.02  149.000000      94.551852  80.000000  7.460000   \n",
+       "\n",
+       "           CardiovascularDisease  RespiratoryFailure  \n",
+       "PatientID                                             \n",
+       "P_131                        0.0                -1.0  \n",
+       "P_132                        0.0                -1.0  \n",
+       "P_195                        1.0                -1.0  \n",
+       "P_193                        0.0                -1.0  \n",
+       "P_140                        0.0                -1.0  \n",
+       "...                          ...                 ...  \n",
+       "P_1_12                       0.0                 0.0  \n",
+       "P_1_8                        0.0                 0.0  \n",
+       "P_1_10                       0.0                 0.0  \n",
+       "P_1_26                       0.0                 0.0  \n",
+       "P_1_146                      1.0                 0.0  \n",
+       "\n",
+       "[863 rows x 13 columns]"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "label_col = \"Prognosis\"\n",
+    "\n",
+    "y = df_reduced[[label_col]]\n",
+    "X_df = df_reduced[[c for c in df_reduced.columns if c not in [label_col]]]\n",
+    "\n",
+    "X_df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "hollow-match",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "picked_features = X_df.columns\n",
+    "\n",
+    "#picked_features = [\n",
+    "#    \"LDH\",\n",
+    "#    \"Age\",\n",
+    "#    \"PaO2\",\n",
+    "#    \"CRP\",\n",
+    "#    \"WBC\",\n",
+    "#    \"pH\"\n",
+    "#]\n",
+    "\n",
+    "#\n",
+    "X = X_df[picked_features]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "38c4a8cc",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# now also load the test data for which maksim imputed the values\n",
+    "\n",
+    "test_path = pathlib.Path(\"/home/starke88/data/haicu/covid_data_challenge_2021/testSet_filled2.txt\")\n",
+    "df_test = pd.read_csv(test_path)\n",
+    "df_test.set_index(\"PatientID\", inplace=True)\n",
+    "#df_test.drop([\"Unnamed: 0\"], axis=1, inplace=True)\n",
+    "# get rid of the features we dumped from the training set as well\n",
+    "df_test.drop([c for c in df_test.columns if c not in df_reduced.columns], axis=1, inplace=True)\n",
+    "# one-hot encode the hospital\n",
+    "#foo = pd.get_dummies(df_test[\"Hospital\"], prefix=\"Hospital\")\n",
+    "#df_test = pd.concat([df_test, foo], axis=1)\n",
+    "#df_test = df_test.drop([\"Hospital\"], axis=1)\n",
+    "assert sorted(df_test.columns) == sorted(X.columns)\n",
+    "\n",
+    "\n",
+    "X_test = df_test"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "7c989ecb",
+   "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>Prognosis</th>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>PatientID</th>\n",
+       "      <th></th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>P_102.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_117.png</th>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_16.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_118.png</th>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_114.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_88.png</th>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_92.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_86.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_9.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>P_90.png</th>\n",
+       "      <td>1</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>120 rows × 1 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "           Prognosis\n",
+       "PatientID           \n",
+       "P_102.png          1\n",
+       "P_117.png          0\n",
+       "P_16.png           1\n",
+       "P_118.png          0\n",
+       "P_114.png          1\n",
+       "...              ...\n",
+       "P_88.png           0\n",
+       "P_92.png           1\n",
+       "P_86.png           1\n",
+       "P_9.png            1\n",
+       "P_90.png           1\n",
+       "\n",
+       "[120 rows x 1 columns]"
+      ]
+     },
+     "execution_count": 11,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "y_test = pd.read_csv(pathlib.Path(\"/home/starke88/data/haicu/covid_data_challenge_2021/solutionTestSet.txt\"))\n",
+    "y_test.set_index(\"ImageFile\", inplace=True)\n",
+    "y_test.index.rename(\"PatientID\", inplace=True)\n",
+    "y_test[\"Prognosis\"] = (y_test[\"Prognosis\"] == \"SEVERE\").astype(np.uint8)\n",
+    "y_test = y_test[[\"Prognosis\"]]\n",
+    "\n",
+    "y_test"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "placed-ratio",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from sklearn.preprocessing import MinMaxScaler, StandardScaler"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "applicable-white",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "['P_384', 'P_417', 'P_551', 'P_479', 'P_555', 'P_376', 'P_1_26', 'P_379', 'P_441', 'P_450', 'P_760', 'P_756', 'P_435', 'P_442', 'P_766', 'P_695', 'P_537', 'P_687', 'P_467', 'P_1_126', 'P_1_75', 'P_653', 'P_1_149', 'P_411', 'P_358', 'P_740', 'P_1_81', 'P_634', 'P_616', 'P_382', 'P_342', 'P_519', 'P_476', 'P_639', 'P_759', 'P_350', 'P_484', 'P_635', 'P_507', 'P_439', 'P_798', 'P_598', 'P_414', 'P_583', 'P_590', 'P_485', 'P_511', 'P_426', 'P_304', 'P_842', 'P_797', 'P_837', 'P_314', 'P_572', 'P_641', 'P_803', 'P_1_55', 'P_702', 'P_691', 'P_518', 'P_1_95', 'P_521', 'P_589', 'P_840', 'P_1_131', 'P_772', 'P_368', 'P_292', 'P_750', 'P_274', 'P_1_146', 'P_650', 'P_505', 'P_733', 'P_623', 'P_697', 'P_620', 'P_1_140', 'P_609', 'P_808', 'P_526', 'P_383', 'P_595', 'P_587', 'P_580', 'P_502', 'P_1_22', 'P_324', 'P_1_72', 'P_725', 'P_336', 'P_491', 'P_781', 'P_407', 'P_546', 'P_645', 'P_425', 'P_735', 'P_445', 'P_812', 'P_385', 'P_1_66', 'P_743', 'P_560', 'P_633']\n",
+      "105\n",
+      "758\n"
+     ]
+    }
+   ],
+   "source": [
+    "valid = list(pd.read_csv(\"/home/starke88/git/covid_data_challenge/clinical_model/valid_ids.csv\", header=None).to_numpy().squeeze())\n",
+    "print(valid)\n",
+    "\n",
+    "print(len(valid))\n",
+    "\n",
+    "\n",
+    "# TODO: train patients are all the others\n",
+    "train = [i for i in X_df.index if i not in valid]\n",
+    "print(len(train))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "convertible-function",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "(758, 13) (758, 1)\n",
+      "(105, 13) (105, 1)\n"
+     ]
+    }
+   ],
+   "source": [
+    "mask_train = X.index.isin(train)\n",
+    "feat_names = X.columns\n",
+    "\n",
+    "X_train = X[mask_train]\n",
+    "y_train = y[mask_train]\n",
+    "\n",
+    "X_valid = X[~mask_train]\n",
+    "y_valid = y[~mask_train]\n",
+    "\n",
+    "print(X_train.shape, y_train.shape)\n",
+    "print(X_valid.shape, y_valid.shape)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "posted-mailing",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "Age                      float64\n",
+       "Sex                      float64\n",
+       "Temp_C                   float64\n",
+       "Cough                    float64\n",
+       "DifficultyInBreathing    float64\n",
+       "WBC                      float64\n",
+       "CRP                      float64\n",
+       "LDH                      float64\n",
+       "Ox_percentage            float64\n",
+       "PaO2                     float64\n",
+       "pH                       float64\n",
+       "CardiovascularDisease    float64\n",
+       "RespiratoryFailure       float64\n",
+       "dtype: object"
+      ]
+     },
+     "execution_count": 15,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "X_train.dtypes"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "altered-payday",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def get_patient_ids(df):\n",
+    "    return df.index.to_numpy().squeeze().tolist()\n",
+    "\n",
+    "scaler = MinMaxScaler()\n",
+    "#scaler = StandardScaler()\n",
+    "\n",
+    "ids_train = get_patient_ids(X_train)\n",
+    "X_train = scaler.fit_transform(X_train.to_numpy())\n",
+    "y_train = y_train.to_numpy().squeeze()\n",
+    "\n",
+    "ids_valid = get_patient_ids(X_valid)\n",
+    "X_valid = scaler.transform(X_valid.to_numpy())\n",
+    "y_valid = y_valid.to_numpy().squeeze()\n",
+    "\n",
+    "ids_test = get_patient_ids(X_test)\n",
+    "X_test = scaler.transform(X_test.to_numpy())\n",
+    "y_test = y_test.to_numpy().squeeze()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "stuffed-connecticut",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "(758, 13) (758,) 758\n",
+      "(105, 13) (105,) 105\n",
+      "(120, 13) (120,) 120\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(X_train.shape, y_train.shape, len(ids_train))\n",
+    "print(X_valid.shape, y_valid.shape, len(ids_valid))\n",
+    "print(X_test.shape, y_test.shape, len(ids_test))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "instant-seating",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{'Age': 2.6797615096744147, 'Sex': -0.6752236739227908, 'Temp_C': 0.8531413177461858, 'Cough': -0.08022687494637566, 'DifficultyInBreathing': 0.9606509001177268, 'WBC': 0.9440718078248774, 'CRP': 4.435017116517157, 'LDH': 22.117162047581637, 'Ox_percentage': -4.870467925676456, 'PaO2': 1.5059465620389225, 'pH': -2.957910963613239, 'CardiovascularDisease': -0.0255222153020973, 'RespiratoryFailure': 2.067007555935972}\n",
+      "\n",
+      "significant coeffs {'Age': 2.6797615096744147, 'Sex': -0.6752236739227908, 'Temp_C': 0.8531413177461858, 'Cough': -0.08022687494637566, 'DifficultyInBreathing': 0.9606509001177268, 'WBC': 0.9440718078248774, 'CRP': 4.435017116517157, 'LDH': 22.117162047581637, 'Ox_percentage': -4.870467925676456, 'PaO2': 1.5059465620389225, 'pH': -2.957910963613239, 'CardiovascularDisease': -0.0255222153020973, 'RespiratoryFailure': 2.067007555935972}\n",
+      "\n",
+      "Training performance\n",
+      "=====\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.78      0.80      0.79       384\n",
+      "           1       0.79      0.77      0.78       374\n",
+      "\n",
+      "    accuracy                           0.78       758\n",
+      "   macro avg       0.78      0.78      0.78       758\n",
+      "weighted avg       0.78      0.78      0.78       758\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[306  78]\n",
+      " [ 86 288]]\n",
+      "\n",
+      "-AUC 0.7834642379679144\n",
+      "\n",
+      "Validation performance\n",
+      "=====\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.83      0.80      0.81        54\n",
+      "           1       0.79      0.82      0.81        51\n",
+      "\n",
+      "    accuracy                           0.81       105\n",
+      "   macro avg       0.81      0.81      0.81       105\n",
+      "weighted avg       0.81      0.81      0.81       105\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[43 11]\n",
+      " [ 9 42]]\n",
+      "\n",
+      "-AUC 0.8099128540305011\n",
+      "\n",
+      "Test performance\n",
+      "=====\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.41      0.91      0.57        35\n",
+      "           1       0.93      0.46      0.61        85\n",
+      "\n",
+      "    accuracy                           0.59       120\n",
+      "   macro avg       0.67      0.69      0.59       120\n",
+      "weighted avg       0.78      0.59      0.60       120\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[32  3]\n",
+      " [46 39]]\n",
+      "\n",
+      "-AUC 0.6865546218487394\n"
+     ]
+    }
+   ],
+   "source": [
+    "## simple logistic regression\n",
+    "from sklearn.linear_model import LogisticRegression\n",
+    "from sklearn.ensemble import RandomForestClassifier\n",
+    "from sklearn.metrics import classification_report, roc_auc_score, roc_curve, confusion_matrix\n",
+    "\n",
+    "#model = RandomForestClassifier() \n",
+    "model = LogisticRegression(penalty=\"none\", solver=\"lbfgs\")\n",
+    "\n",
+    "model.fit(X_train, y_train)\n",
+    "if hasattr(model, \"coef_\"):\n",
+    "    coeffs = dict(zip(feat_names, model.coef_[0]))\n",
+    "    print(coeffs)\n",
+    "    print(\"\\nsignificant coeffs\", {k:v for k,v in coeffs.items() if np.abs(v) > .01})\n",
+    "print()\n",
+    "\n",
+    "def evaluate_model(model, X, y):\n",
+    "    pred_y = model.predict(X)\n",
+    "\n",
+    "    print(\"-Classification metrics\\n\", classification_report(y, pred_y))\n",
+    "    print()\n",
+    "    print(\"-Confusion matrix\\n\", confusion_matrix(y, pred_y))\n",
+    "    print()\n",
+    "    print(\"-AUC\", roc_auc_score(y, pred_y))\n",
+    "    \n",
+    "\n",
+    "def make_prediction_df(model, X, patient_ids, method_name):\n",
+    "    cls_1_index = np.where(model.classes_ == 1)[0][0]\n",
+    "    pred_score = model.predict_proba(X)[:, cls_1_index]\n",
+    "    return pd.DataFrame({\"PatientID\": patient_ids, f\"prediction_{method_name}\": pred_score})\n",
+    "    \n",
+    "    \n",
+    "print(\"Training performance\\n=====\")\n",
+    "evaluate_model(model, X_train, y_train)\n",
+    "\n",
+    "print(\"\\nValidation performance\\n=====\")\n",
+    "evaluate_model(model, X_valid, y_valid)\n",
+    "\n",
+    "print(\"\\nTest performance\\n=====\")\n",
+    "evaluate_model(model, X_test, y_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "a7ceb785",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model_name = \"Clinical-logistic-regression-nopenalty-lbfgs-trainedOn758PatientsWithImputation\"\n",
+    "pred_train = make_prediction_df(model, X_train, patient_ids=ids_train, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_valid = make_prediction_df(model, X_valid, patient_ids=ids_valid, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_test = make_prediction_df(model, X_test, patient_ids=ids_test, \n",
+    "                   method_name=model_name)\n",
+    "pred_test.to_csv(f\"/home/starke88/git/covid_data_challenge/clinical_model/predictions_{model_name}.csv\", index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "chemical-organ",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "if hasattr(model, \"feature_importances_\"):\n",
+    "    imp = dict(zip(feat_names, model.feature_importances_))\n",
+    "    for k, v in imp.items():\n",
+    "        print(k,v)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "scientific-cancer",
+   "metadata": {
+    "scrolled": true
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Fitting 5 folds for each of 4 candidates, totalling 20 fits\n",
+      ": grid-search best parameters {'fit__class_weight': 'balanced', 'fit__penalty': 'l1'}\n",
+      ": grid-search best score 0.7559515510630882\n",
+      "\n",
+      "\n",
+      "Training performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.76      0.78      0.77       384\n",
+      "           1       0.77      0.75      0.76       374\n",
+      "\n",
+      "    accuracy                           0.77       758\n",
+      "   macro avg       0.77      0.77      0.77       758\n",
+      "weighted avg       0.77      0.77      0.77       758\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[300  84]\n",
+      " [ 93 281]]\n",
+      "\n",
+      "-AUC 0.766293449197861\n",
+      "\n",
+      "Validation performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.83      0.80      0.81        54\n",
+      "           1       0.79      0.82      0.81        51\n",
+      "\n",
+      "    accuracy                           0.81       105\n",
+      "   macro avg       0.81      0.81      0.81       105\n",
+      "weighted avg       0.81      0.81      0.81       105\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[43 11]\n",
+      " [ 9 42]]\n",
+      "\n",
+      "-AUC 0.8099128540305011\n",
+      "\n",
+      "Test performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.42      0.91      0.57        35\n",
+      "           1       0.93      0.47      0.62        85\n",
+      "\n",
+      "    accuracy                           0.60       120\n",
+      "   macro avg       0.67      0.69      0.60       120\n",
+      "weighted avg       0.78      0.60      0.61       120\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[32  3]\n",
+      " [45 40]]\n",
+      "\n",
+      "-AUC 0.6924369747899161\n"
+     ]
+    }
+   ],
+   "source": [
+    "from sklearn.pipeline import Pipeline\n",
+    "from sklearn.model_selection import GridSearchCV, KFold\n",
+    "\n",
+    "# l1 or l2\n",
+    "logistic_pipe = Pipeline([\n",
+    "    #(\"scaling\", MinMaxScaler()),\n",
+    "    (\"fit\", LogisticRegression(solver=\"liblinear\"))])\n",
+    "\n",
+    "logistic_grid = {\n",
+    "    #\"scaling\": [\"passthrough\", MinMaxScaler(), StandardScaler()],\n",
+    "    \"fit__penalty\": [\"l1\", \"l2\"],\n",
+    "    \"fit__class_weight\": [None, \"balanced\"]}\n",
+    "\n",
+    "\n",
+    "searcher = GridSearchCV(\n",
+    "            estimator=logistic_pipe,\n",
+    "            param_grid=logistic_grid,\n",
+    "            cv=KFold(n_splits=5, shuffle=False),\n",
+    "            #scoring=\"neg_mean_absolute_error\",\n",
+    "            verbose=1)\n",
+    "searcher.fit(X_train, y_train)\n",
+    "print(\": grid-search best parameters\", searcher.best_params_)\n",
+    "print(\": grid-search best score\", searcher.best_score_)\n",
+    "print()\n",
+    "print(\"\\nTraining performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_train, y_train)\n",
+    "print(\"\\nValidation performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_valid, y_valid)\n",
+    "print(\"\\nTest performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_test, y_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "id": "a499fb59",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model_name = \"Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn758PatientsWithImputation\"\n",
+    "model = searcher.best_estimator_\n",
+    "\n",
+    "pred_train = make_prediction_df(model, X_train, patient_ids=ids_train, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_valid = make_prediction_df(model, X_valid, patient_ids=ids_valid, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_test = make_prediction_df(model, X_test, patient_ids=ids_test, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_test.to_csv(f\"/home/starke88/git/covid_data_challenge/clinical_model/predictions_{model_name}.csv\", index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "dominican-colonial",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Fitting 5 folds for each of 400 candidates, totalling 2000 fits\n",
+      ": grid-search best parameters {'fit__max_depth': 20, 'fit__max_leaf_nodes': 100, 'fit__min_impurity_decrease': 0, 'fit__n_estimators': 100}\n",
+      ": grid-search best score 0.7533809689787383\n",
+      "\n",
+      "\n",
+      "Training performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.99      0.99      0.99       384\n",
+      "           1       0.99      0.99      0.99       374\n",
+      "\n",
+      "    accuracy                           0.99       758\n",
+      "   macro avg       0.99      0.99      0.99       758\n",
+      "weighted avg       0.99      0.99      0.99       758\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[379   5]\n",
+      " [  3 371]]\n",
+      "\n",
+      "-AUC 0.9894788881461675\n",
+      "\n",
+      "Validation performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.79      0.78      0.79        54\n",
+      "           1       0.77      0.78      0.78        51\n",
+      "\n",
+      "    accuracy                           0.78       105\n",
+      "   macro avg       0.78      0.78      0.78       105\n",
+      "weighted avg       0.78      0.78      0.78       105\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[42 12]\n",
+      " [11 40]]\n",
+      "\n",
+      "-AUC 0.7810457516339868\n",
+      "\n",
+      "Test performance of best model\n",
+      "\n",
+      "-Classification metrics\n",
+      "               precision    recall  f1-score   support\n",
+      "\n",
+      "           0       0.44      0.91      0.59        35\n",
+      "           1       0.94      0.52      0.67        85\n",
+      "\n",
+      "    accuracy                           0.63       120\n",
+      "   macro avg       0.69      0.72      0.63       120\n",
+      "weighted avg       0.79      0.63      0.65       120\n",
+      "\n",
+      "\n",
+      "-Confusion matrix\n",
+      " [[32  3]\n",
+      " [41 44]]\n",
+      "\n",
+      "-AUC 0.7159663865546217\n"
+     ]
+    }
+   ],
+   "source": [
+    "rf_pipe = Pipeline([\n",
+    "    #(\"scaling\", MinMaxScaler()),\n",
+    "    (\"fit\", RandomForestClassifier(n_estimators=100,\n",
+    "                                  # max_depth=5,\n",
+    "                                  # max_leaf_nodes=20,\n",
+    "                                  # min_impurity_decrease=.1\n",
+    "                                  random_state=42,\n",
+    "                                  ))\n",
+    "])\n",
+    "rf_grid = {\n",
+    "    #\"scaling\": [MinMaxScaler(), StandardScaler()],\n",
+    "    \"fit__n_estimators\": [5, 10, 20, 100, 200],\n",
+    "    \"fit__max_depth\": [None, 5, 10, 20, 50],\n",
+    "    \"fit__max_leaf_nodes\": [None, 20, 50, 100],\n",
+    "    \"fit__min_impurity_decrease\": [0, 0.1, .5, 1]    }\n",
+    "\n",
+    "searcher = GridSearchCV(\n",
+    "            estimator=rf_pipe,\n",
+    "            param_grid=rf_grid,\n",
+    "            cv=KFold(n_splits=5, shuffle=False),\n",
+    "            #scoring=\"neg_mean_absolute_error\",\n",
+    "            verbose=1)\n",
+    "searcher.fit(X_train, y_train)\n",
+    "\n",
+    "print(\": grid-search best parameters\", searcher.best_params_)\n",
+    "print(\": grid-search best score\", searcher.best_score_)\n",
+    "print()\n",
+    "print(\"\\nTraining performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_train, y_train)\n",
+    "print(\"\\nValidation performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_valid, y_valid)\n",
+    "print(\"\\nTest performance of best model\\n\")\n",
+    "evaluate_model(searcher.best_estimator_, X_test, y_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "id": "taken-mortgage",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model_name = \"Clinical-randomforest-100estimators-trainedOn758PatientsWithImputation\"\n",
+    "model = searcher.best_estimator_\n",
+    "\n",
+    "pred_train = make_prediction_df(model, X_train, patient_ids=ids_train, \n",
+    "                   method_name=model_name)\n",
+    "\n",
+    "pred_valid = make_prediction_df(model, X_valid, patient_ids=ids_valid, \n",
+    "                   method_name=model_name)\n",
+    "pred_test = make_prediction_df(model, X_test, patient_ids=ids_test, \n",
+    "                   method_name=model_name)\n",
+    "pred_test.to_csv(f\"/home/starke88/git/covid_data_challenge/clinical_model/predictions_{model_name}.csv\", index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a2163966",
+   "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": "3.8.8"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
diff --git a/clinical_model/predictions_Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn312Patients.csv b/clinical_model/predictions_Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn312Patients.csv
new file mode 100644
index 0000000..aeedc33
--- /dev/null
+++ b/clinical_model/predictions_Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn312Patients.csv
@@ -0,0 +1,121 @@
+PatientID,prediction_Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn312Patients
+P_102,0.33520004647784835
+P_117,0.06827689247042472
+P_16,0.13584828184647155
+P_118,0.7975236836311405
+P_114,0.4282911929313953
+P_2,0.5607635032130464
+P_20,0.8791060732905628
+P_26,0.37615970263923376
+P_38,0.17539619031427528
+P_41,0.5662134515078459
+P_49,0.1607696261695052
+P_45,0.39998307827296087
+P_40,0.6567350933745325
+P_55,0.15078852023168587
+P_58,0.6936727399468227
+P_53,0.07938684251078379
+P_54,0.18994299016175986
+P_65,0.2471092057157912
+P_68,0.21118541518519335
+P_7,0.07290098155043286
+P_75,0.2270635764561114
+P_112,0.5377720346883706
+P_116,0.19923650965029246
+P_14,0.5841029589439515
+P_17,0.764786127124441
+P_111,0.44908579731238213
+P_1,0.17901755433065009
+P_106,0.7724132167599
+P_113,0.45708141399737634
+P_104,0.3294183049598339
+P_21,0.41069691477025
+P_25,0.3512817849966773
+P_44,0.08860785830460742
+P_46,0.27606226883431606
+P_43,0.20652410772080362
+P_30,0.6120256368189827
+P_50,0.41241659393951874
+P_60,0.3891833520344404
+P_6,0.4580981480214153
+P_69,0.18005363844199188
+P_66,0.11480638012803182
+P_85,0.8581273726962804
+P_94,0.2278812627780469
+P_84,0.18295590526231506
+P_83,0.12304522933578058
+P_91,0.24543700816344285
+P_76,0.2198964059213189
+P_77,0.530927099559969
+P_81,0.4658834530602529
+P_82,0.8765728099098288
+P_93,0.2307890949678067
+P_98,0.2817343412587218
+P_96,0.3128442272142378
+P_95,0.420416680576292
+P_80,0.875917955810806
+P_87,0.8473339962284742
+P_11,0.12939075660289623
+P_115,0.17932515301714272
+P_18,0.3412197166013632
+P_12,0.4085514105511912
+P_119,0.31999979532889444
+P_15,0.3317820046672228
+P_100,0.442321181987111
+P_107,0.616294812707784
+P_10,0.9886770584328072
+P_101,0.7923609634624441
+P_108,0.08003254948363757
+P_110,0.8563301321803212
+P_24,0.665118287186855
+P_28,0.8138245654764317
+P_23,0.8063273894728695
+P_29,0.3245450002343199
+P_32,0.606421411801071
+P_37,0.6802529357544478
+P_34,0.7860112241050077
+P_35,0.626668308143247
+P_36,0.34647149057349647
+P_39,0.09921526941333601
+P_4,0.7419867041541264
+P_42,0.376626436080426
+P_63,0.5375326047740958
+P_61,0.18159398036610488
+P_62,0.4121356368846893
+P_59,0.4139631148649756
+P_70,0.32800228129546755
+P_72,0.363302970875322
+P_74,0.22693675422272985
+P_97,0.4503424328184856
+P_99,0.21009202275843925
+P_73,0.19939571075463042
+P_78,0.18419547543075346
+P_71,0.3254840991389562
+P_120,0.3456758805905774
+P_13,0.3002468351439
+P_22,0.5961214844455944
+P_105,0.7083471027674944
+P_109,0.6882021260968174
+P_103,0.4777795011847989
+P_19,0.25850371254408455
+P_27,0.26326323331807183
+P_3,0.34560859025267243
+P_33,0.12648624250858967
+P_31,0.7299946886551383
+P_47,0.46315094593381667
+P_48,0.20489716776284475
+P_57,0.16939842591837587
+P_5,0.2860990510662038
+P_51,0.8770743465243397
+P_56,0.19027433629860382
+P_52,0.08535363595664093
+P_64,0.4535418761626194
+P_67,0.37293363879865526
+P_79,0.3726676202504323
+P_8,0.3187365182964029
+P_89,0.3424301683759125
+P_88,0.20473068036960596
+P_92,0.05807647412335641
+P_86,0.4948297380918963
+P_9,0.6778152753642893
+P_90,0.32128617511339685
diff --git a/clinical_model/predictions_Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn758PatientsWithImputation.csv b/clinical_model/predictions_Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn758PatientsWithImputation.csv
new file mode 100644
index 0000000..359f7be
--- /dev/null
+++ b/clinical_model/predictions_Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn758PatientsWithImputation.csv
@@ -0,0 +1,121 @@
+PatientID,prediction_Clinical-logistic-regression-gridsearchedParams-l1penalty-liblinear-trainedOn758PatientsWithImputation
+P_102,0.3859644918622136
+P_117,0.07284203230498752
+P_16,0.07596895381275044
+P_118,0.9274230083127741
+P_114,0.5142052481657664
+P_2,0.5203429898535421
+P_20,0.8000266412980864
+P_26,0.22480692845410072
+P_38,0.1822518446772093
+P_41,0.5716081459889104
+P_49,0.2473877082260779
+P_45,0.16565900991729404
+P_40,0.7093868803545182
+P_55,0.1979382554661376
+P_58,0.9370410730987137
+P_53,0.08148851451642634
+P_54,0.08294504685675824
+P_65,0.37617956013175763
+P_68,0.35082219144166366
+P_7,0.12534370993786753
+P_75,0.19789852978823735
+P_112,0.6707704472783014
+P_116,0.19805373425310113
+P_14,0.6263954320158809
+P_17,0.5513416542968695
+P_111,0.428343994858313
+P_1,0.13438014802808462
+P_106,0.7783368402394052
+P_113,0.31689459434644496
+P_104,0.24976073218044986
+P_21,0.23995234648296687
+P_25,0.22963990332329398
+P_44,0.09021678959182249
+P_46,0.3781880682643061
+P_43,0.3265208524727368
+P_30,0.6997096980573042
+P_50,0.4426652399472462
+P_60,0.3207795092231149
+P_6,0.1934011763186107
+P_69,0.13343361173621096
+P_66,0.0740632399118832
+P_85,0.8854477880159981
+P_94,0.23190371268296364
+P_84,0.1955168690975254
+P_83,0.23803926535577632
+P_91,0.23848790972929118
+P_76,0.2843177940396236
+P_77,0.8673916935008623
+P_81,0.4004369413550231
+P_82,0.8742958150342303
+P_93,0.08080227666418292
+P_98,0.467161058965024
+P_96,0.23565533294391522
+P_95,0.5445814366451247
+P_80,0.8282874671584735
+P_87,0.9555111563767205
+P_11,0.059184061740297486
+P_115,0.3225233207342592
+P_18,0.2550110459899284
+P_12,0.623587242858893
+P_119,0.49512849763874905
+P_15,0.5608572003946439
+P_100,0.21366138277655405
+P_107,0.7598771587425981
+P_10,0.9919207015221694
+P_101,0.7063801922502047
+P_108,0.18363395680077926
+P_110,0.9421521314646447
+P_24,0.7365033803070948
+P_28,0.7818310300697472
+P_23,0.6808312092485941
+P_29,0.6400896202911683
+P_32,0.44742023295516
+P_37,0.6946437690664641
+P_34,0.7681504496611707
+P_35,0.600524125260131
+P_36,0.3665075968331374
+P_39,0.17383317531486894
+P_4,0.8587062403466065
+P_42,0.3232414000393499
+P_63,0.7475300485626806
+P_61,0.1767042091824176
+P_62,0.46973162577547806
+P_59,0.43752823467747265
+P_70,0.19947511560093817
+P_72,0.3100108707372959
+P_74,0.1443317377442051
+P_97,0.295647292893258
+P_99,0.2170817253904323
+P_73,0.20278616364162716
+P_78,0.18479805444077033
+P_71,0.581513813023997
+P_120,0.34810997392105003
+P_13,0.5975978103825967
+P_22,0.5369159972769143
+P_105,0.63716075585888
+P_109,0.6121339462956245
+P_103,0.30342540538527224
+P_19,0.10225471287027968
+P_27,0.1726493557480162
+P_3,0.2404355338123172
+P_33,0.31041182111727883
+P_31,0.835361706924908
+P_47,0.3810059842730029
+P_48,0.27986321930047503
+P_57,0.46219508350777755
+P_5,0.08958116840252862
+P_51,0.9275899520989356
+P_56,0.13710198371029936
+P_52,0.07490890194571004
+P_64,0.569648866590487
+P_67,0.4214220485937342
+P_79,0.2721423316936425
+P_8,0.3092079108946637
+P_89,0.26971439193936825
+P_88,0.20075491283355626
+P_92,0.07237953993542133
+P_86,0.5736632700761363
+P_9,0.6491348300524976
+P_90,0.28817196014750335
diff --git a/clinical_model/predictions_Clinical-logistic-regression-nopenalty-lbfgs-trainedOn312Patients.csv b/clinical_model/predictions_Clinical-logistic-regression-nopenalty-lbfgs-trainedOn312Patients.csv
new file mode 100644
index 0000000..386810a
--- /dev/null
+++ b/clinical_model/predictions_Clinical-logistic-regression-nopenalty-lbfgs-trainedOn312Patients.csv
@@ -0,0 +1,121 @@
+PatientID,prediction_Clinical-logistic-regression-nopenalty-lbfgs-trainedOn312Patients
+P_102,0.24903656599446405
+P_117,0.04003292783510478
+P_16,0.12446757810452885
+P_118,0.9218160374499078
+P_114,0.4286919261325882
+P_2,0.6640975112163694
+P_20,0.9701462449738542
+P_26,0.40350849858638743
+P_38,0.1208567522913325
+P_41,0.7725482931492166
+P_49,0.07875418683872765
+P_45,0.4343138006325572
+P_40,0.7276851679444762
+P_55,0.15253447403064463
+P_58,0.8784259011286092
+P_53,0.025821151017420253
+P_54,0.23949971971236725
+P_65,0.12741975197440444
+P_68,0.20367678616280735
+P_7,0.02510497766076813
+P_75,0.10381775699916347
+P_112,0.5862375095444788
+P_116,0.1541375635924125
+P_14,0.7447034500619543
+P_17,0.8043590491041641
+P_111,0.3928752589840098
+P_1,0.13477245574194524
+P_106,0.8036419967131622
+P_113,0.2999517103060956
+P_104,0.3183387823212139
+P_21,0.5734787382187158
+P_25,0.24574649024284842
+P_44,0.0370316065900205
+P_46,0.22916173088150432
+P_43,0.12339156681362115
+P_30,0.6928260601737747
+P_50,0.32724604748247815
+P_60,0.48186534037452167
+P_6,0.5095930829252183
+P_69,0.1421954988590595
+P_66,0.08087919589605003
+P_85,0.9022399978238933
+P_94,0.10188391243468453
+P_84,0.24780992174710426
+P_83,0.07991659330774005
+P_91,0.2181067932193218
+P_76,0.12944253336885472
+P_77,0.41860703749776523
+P_81,0.4749830682794505
+P_82,0.9321785218637345
+P_93,0.23557437027348954
+P_98,0.369719152762019
+P_96,0.25844036130117654
+P_95,0.28040712453966216
+P_80,0.9771688451492363
+P_87,0.8363211321161775
+P_11,0.12495360676520206
+P_115,0.10173593582734358
+P_18,0.22990816675523057
+P_12,0.38342801146690786
+P_119,0.3162849081476485
+P_15,0.2377489248825254
+P_100,0.5675477509654552
+P_107,0.7246437743095745
+P_10,0.9994069351568323
+P_101,0.8971340337656126
+P_108,0.04077080591014789
+P_110,0.8751943820848411
+P_24,0.8781596323416284
+P_28,0.9644450441403449
+P_23,0.9309188166665078
+P_29,0.3510101876751694
+P_32,0.7051648055228926
+P_37,0.6414231072570596
+P_34,0.882659254222237
+P_35,0.7156817381482089
+P_36,0.28832412541491176
+P_39,0.030598088898500145
+P_4,0.8263203337126278
+P_42,0.297984514254678
+P_63,0.6694741987477717
+P_61,0.09583497157167596
+P_62,0.29356760542853205
+P_59,0.3500394829880046
+P_70,0.2890625180939771
+P_72,0.4674465959129291
+P_74,0.15504646053445142
+P_97,0.43165048417498425
+P_99,0.11431370256706168
+P_73,0.11272392805569711
+P_78,0.11492311015642452
+P_71,0.21199110034589025
+P_120,0.27036557823277135
+P_13,0.21381100516035592
+P_22,0.5516303971242341
+P_105,0.8390841609161274
+P_109,0.7820717005086841
+P_103,0.41686027760095956
+P_19,0.2729547667960578
+P_27,0.2537375276543555
+P_3,0.2738278688121344
+P_33,0.07477187752443443
+P_31,0.8448419680451451
+P_47,0.43146175916292073
+P_48,0.15466418064258186
+P_57,0.11060379553685333
+P_5,0.3522418049485151
+P_51,0.9279976991287591
+P_56,0.13622296469253095
+P_52,0.054660053928964995
+P_64,0.7182330795665481
+P_67,0.2767831256918614
+P_79,0.5145329801340882
+P_8,0.2647806566559488
+P_89,0.3273118637511314
+P_88,0.09156193298298344
+P_92,0.02644416054907079
+P_86,0.5000568945321757
+P_9,0.8194760027662711
+P_90,0.2293672292636861
diff --git a/clinical_model/predictions_Clinical-logistic-regression-nopenalty-lbfgs-trainedOn758PatientsWithImputation.csv b/clinical_model/predictions_Clinical-logistic-regression-nopenalty-lbfgs-trainedOn758PatientsWithImputation.csv
new file mode 100644
index 0000000..6572a5d
--- /dev/null
+++ b/clinical_model/predictions_Clinical-logistic-regression-nopenalty-lbfgs-trainedOn758PatientsWithImputation.csv
@@ -0,0 +1,121 @@
+PatientID,prediction_Clinical-logistic-regression-nopenalty-lbfgs-trainedOn758PatientsWithImputation
+P_102,0.31835132661122767
+P_117,0.031654828816397126
+P_16,0.03742451771059102
+P_118,0.9263724726367898
+P_114,0.4487215180364961
+P_2,0.41209183762170254
+P_20,0.9413581709090953
+P_26,0.12254455868162095
+P_38,0.13686780719998914
+P_41,0.5546815828055516
+P_49,0.1853769046994918
+P_45,0.10769026257611337
+P_40,0.6935879208992427
+P_55,0.12311850411956264
+P_58,0.901061569837301
+P_53,0.05130167506402003
+P_54,0.038654491665258334
+P_65,0.2787366639095199
+P_68,0.2835908781423644
+P_7,0.06928335619636614
+P_75,0.18594873175266755
+P_112,0.6715641643206077
+P_116,0.16121286586660563
+P_14,0.6837813965299712
+P_17,0.6570071771667383
+P_111,0.47371976332323207
+P_1,0.07678311041218151
+P_106,0.8314763317530923
+P_113,0.3128436911901391
+P_104,0.2676081896898872
+P_21,0.1782253660252437
+P_25,0.20243887091158302
+P_44,0.059795204792418614
+P_46,0.30253807654342935
+P_43,0.3170826315702827
+P_30,0.7659238004605401
+P_50,0.453595275989049
+P_60,0.18855187541682203
+P_6,0.13359181337363316
+P_69,0.07185082039443393
+P_66,0.036344064257934225
+P_85,0.9014752010278936
+P_94,0.19104165975823614
+P_84,0.13230119920477593
+P_83,0.16897721157353496
+P_91,0.22527919584513
+P_76,0.21418197848603898
+P_77,0.8470080841044276
+P_81,0.40078826191809735
+P_82,0.9384661016546725
+P_93,0.04338501174989692
+P_98,0.4292495039269186
+P_96,0.18091404535537223
+P_95,0.51773101630826
+P_80,0.8182041515561647
+P_87,0.9735490678089487
+P_11,0.02846983665068818
+P_115,0.30101989295006626
+P_18,0.20607772228612625
+P_12,0.6211695827540649
+P_119,0.561789957082007
+P_15,0.5594166920975943
+P_100,0.15215523081216822
+P_107,0.7683849853685651
+P_10,0.9990551129397476
+P_101,0.6336624377345574
+P_108,0.10445505154862177
+P_110,0.9676982887548282
+P_24,0.6523241473862897
+P_28,0.8455845387060906
+P_23,0.6945626539230936
+P_29,0.6508184140286792
+P_32,0.5114257954822891
+P_37,0.7066189076650161
+P_34,0.8165771053854749
+P_35,0.6097645210880083
+P_36,0.30071317136038617
+P_39,0.1162753785938526
+P_4,0.8831712394278431
+P_42,0.3237176291753919
+P_63,0.7803444393399233
+P_61,0.13177536731575853
+P_62,0.416799561036709
+P_59,0.38225233800286906
+P_70,0.12265547315914713
+P_72,0.16934677202379259
+P_74,0.12062993425169195
+P_97,0.28002036014426435
+P_99,0.1679643155751617
+P_73,0.1611190550566634
+P_78,0.1664093982438191
+P_71,0.6085060575977724
+P_120,0.48033453626212197
+P_13,0.5701560862816882
+P_22,0.5621676251919606
+P_105,0.7336890101568924
+P_109,0.6593722999411834
+P_103,0.28491705072494317
+P_19,0.057279600804310396
+P_27,0.10588685917484085
+P_3,0.20766057194703635
+P_33,0.24522873540184817
+P_31,0.8179099011904133
+P_47,0.4913721860539052
+P_48,0.2331864252558713
+P_57,0.39122463485552156
+P_5,0.05547535275840547
+P_51,0.9426698156648277
+P_56,0.06723012997180255
+P_52,0.03595983576613419
+P_64,0.4828763552319358
+P_67,0.3941998337156697
+P_79,0.23583730755162813
+P_8,0.2713379872569081
+P_89,0.19709344495463335
+P_88,0.1265176302770373
+P_92,0.029902999142444903
+P_86,0.6033392456414456
+P_9,0.5807802704104573
+P_90,0.23251775468693256
diff --git a/clinical_model/predictions_Clinical-randomforest-100estimators-trainedOn758PatientsWithImputation.csv b/clinical_model/predictions_Clinical-randomforest-100estimators-trainedOn758PatientsWithImputation.csv
new file mode 100644
index 0000000..a1e19ab
--- /dev/null
+++ b/clinical_model/predictions_Clinical-randomforest-100estimators-trainedOn758PatientsWithImputation.csv
@@ -0,0 +1,121 @@
+PatientID,prediction_Clinical-randomforest-100estimators-trainedOn758PatientsWithImputation
+P_102,0.38946550336041985
+P_117,0.21282102831486885
+P_16,0.2292309049178195
+P_118,0.8573664001365144
+P_114,0.457689330355186
+P_2,0.8420358002655038
+P_20,0.7528432827774967
+P_26,0.24425810386334748
+P_38,0.29173251483568535
+P_41,0.656733411843052
+P_49,0.08799841020508709
+P_45,0.06698014714367952
+P_40,0.42709772892278974
+P_55,0.44534673653583484
+P_58,0.5900473936649999
+P_53,0.1530290965640159
+P_54,0.22204535598555325
+P_65,0.571362045176132
+P_68,0.29297598199300995
+P_7,0.2994842037318093
+P_75,0.14094637035751595
+P_112,0.5946715797653012
+P_116,0.07962602409320528
+P_14,0.7026038037522898
+P_17,0.6877395293713917
+P_111,0.384172711100478
+P_1,0.14695364889845955
+P_106,0.6628320711542883
+P_113,0.2468814460870966
+P_104,0.5301858443644191
+P_21,0.24849563982118067
+P_25,0.262311664929914
+P_44,0.027075693448032782
+P_46,0.31094975734390945
+P_43,0.6320940826036633
+P_30,0.8124279049116454
+P_50,0.2837401202915044
+P_60,0.6210785165034133
+P_6,0.1778964334173122
+P_69,0.02551757876760926
+P_66,0.01922233378480941
+P_85,0.7003444699466498
+P_94,0.10053982601769644
+P_84,0.4213463524866704
+P_83,0.4726997704752738
+P_91,0.09917730976213647
+P_76,0.20653495968118804
+P_77,0.582829309993914
+P_81,0.4174469791476917
+P_82,0.7558552935853973
+P_93,0.1558813883923038
+P_98,0.6312581558880244
+P_96,0.2833861091846772
+P_95,0.5094860156965737
+P_80,0.8803020997111964
+P_87,0.798605070433317
+P_11,0.2375383247517553
+P_115,0.4060607417634127
+P_18,0.15221828087832698
+P_12,0.6895586841901787
+P_119,0.7532326563388537
+P_15,0.6162961454854518
+P_100,0.2656280351221656
+P_107,0.6075562025213671
+P_10,0.819287313528472
+P_101,0.5418641730780982
+P_108,0.2065326710717261
+P_110,0.7644121052361565
+P_24,0.7197081996709395
+P_28,0.8517132591972909
+P_23,0.5829919166506827
+P_29,0.852096502982105
+P_32,0.4840101160178048
+P_37,0.5665506224238773
+P_34,0.5382771261139793
+P_35,0.4385345621302919
+P_36,0.05823659162874619
+P_39,0.02019208733288843
+P_4,0.6970987996312197
+P_42,0.2808675912614251
+P_63,0.7566663986230973
+P_61,0.2562584769528749
+P_62,0.5566724226360306
+P_59,0.3504794703875266
+P_70,0.15315224977807163
+P_72,0.20977103991222573
+P_74,0.135401898322289
+P_97,0.05458064834727395
+P_99,0.1466862165111942
+P_73,0.06753047604515297
+P_78,0.3309144651899263
+P_71,0.5088787113144024
+P_120,0.4531747487154755
+P_13,0.6939067244800701
+P_22,0.1438258555795674
+P_105,0.783879444593995
+P_109,0.5463574926544849
+P_103,0.08031589638700326
+P_19,0.1050073216613282
+P_27,0.3458076676894414
+P_3,0.1028003349570211
+P_33,0.4926338529597794
+P_31,0.650329965958054
+P_47,0.5183042333533406
+P_48,0.15527066117798724
+P_57,0.4430172205023816
+P_5,0.05170384323132989
+P_51,0.8645575025560717
+P_56,0.3186072782197584
+P_52,0.203838418651035
+P_64,0.6492725510950499
+P_67,0.07489726418238614
+P_79,0.20621292474964892
+P_8,0.19378498619689222
+P_89,0.34821110062882354
+P_88,0.19459208091786806
+P_92,0.33073754363143565
+P_86,0.6885462291774193
+P_9,0.8487928995809984
+P_90,0.21821745917040292
diff --git a/clinical_model/predictions_Clinical-randomforest-10estimators-trainedOn312Patients.csv b/clinical_model/predictions_Clinical-randomforest-10estimators-trainedOn312Patients.csv
new file mode 100644
index 0000000..216bd16
--- /dev/null
+++ b/clinical_model/predictions_Clinical-randomforest-10estimators-trainedOn312Patients.csv
@@ -0,0 +1,121 @@
+PatientID,prediction_Clinical-randomforest-10estimators-trainedOn312Patients
+P_102,0.6425278613155709
+P_117,0.23339452285392942
+P_16,0.382990132037891
+P_118,0.9282818035426731
+P_114,0.3877030197450563
+P_2,0.7176854742830476
+P_20,0.5874486027316216
+P_26,0.24590858248675715
+P_38,0.19759922660099263
+P_41,0.38828407284289634
+P_49,0.05482132198137636
+P_45,0.12953760353060625
+P_40,0.6554076846741177
+P_55,0.25681802424271527
+P_58,0.6599216988901248
+P_53,0.2728883144616123
+P_54,0.07001806553632886
+P_65,0.8024287344622024
+P_68,0.2804062441448522
+P_7,0.1863494735731443
+P_75,0.24436174441330313
+P_112,0.45450655406537754
+P_116,0.12089480074142227
+P_14,0.731673665061823
+P_17,0.757881631006631
+P_111,0.3625352904213991
+P_1,0.032433262279883815
+P_106,0.600067587054933
+P_113,0.25120229429085184
+P_104,0.5463005309217851
+P_21,0.33596179970040774
+P_25,0.4510787063430278
+P_44,0.032433262279883815
+P_46,0.15610908117428002
+P_43,0.4316392252026581
+P_30,0.5744508172139752
+P_50,0.2103876881394542
+P_60,0.6400735339845401
+P_6,0.11473607048947523
+P_69,0.032433262279883815
+P_66,0.032433262279883815
+P_85,0.49854260257562133
+P_94,0.20359240249774424
+P_84,0.6414821048847262
+P_83,0.2095265121315124
+P_91,0.2746567804301041
+P_76,0.05482132198137636
+P_77,0.8015773594690364
+P_81,0.42645251042634535
+P_82,0.7441847826086957
+P_93,0.06975935198789951
+P_98,0.7455578961017846
+P_96,0.26838604212465017
+P_95,0.2927685582052161
+P_80,0.7342954826321864
+P_87,0.9049493350852046
+P_11,0.14057321786538174
+P_115,0.14558276538480636
+P_18,0.08257265122063609
+P_12,0.9075995288836542
+P_119,0.6860976506630628
+P_15,0.6853565583352818
+P_100,0.42991953856728654
+P_107,0.7551234131021365
+P_10,0.7669739653621233
+P_101,0.8637369482792824
+P_108,0.10063518785885858
+P_110,0.8467355353331086
+P_24,0.7273715316660445
+P_28,0.9045257276507277
+P_23,0.5647038950715422
+P_29,0.5578085982873215
+P_32,0.47887474182724776
+P_37,0.6939094689094689
+P_34,0.657979461695531
+P_35,0.4935015308235123
+P_36,0.05482132198137636
+P_39,0.032433262279883815
+P_4,0.4665488814131521
+P_42,0.33675962351864125
+P_63,0.5857653713115479
+P_61,0.06534763777085004
+P_62,0.6061990228846782
+P_59,0.5056977911592361
+P_70,0.07228525227319468
+P_72,0.340420308960112
+P_74,0.05482132198137636
+P_97,0.05642300842432544
+P_99,0.12089480074142227
+P_73,0.12302324756035113
+P_78,0.5295393413775765
+P_71,0.4766596480201131
+P_120,0.6425225983739916
+P_13,0.6093354146721033
+P_22,0.3162071709183517
+P_105,0.6609344558913878
+P_109,0.1607578369964451
+P_103,0.29633565852841953
+P_19,0.32486360470515396
+P_27,0.155782582555422
+P_3,0.23361362904373592
+P_33,0.47015387670458153
+P_31,0.6582500280892166
+P_47,0.38842939314057395
+P_48,0.14912003634370707
+P_57,0.15576996142033422
+P_5,0.07881106812581798
+P_51,0.9388180967968202
+P_56,0.04989719257170215
+P_52,0.10063518785885858
+P_64,0.6925484526851948
+P_67,0.2087375332746368
+P_79,0.3102554658777222
+P_8,0.2283313033965022
+P_89,0.12390655119871508
+P_88,0.3149504891021572
+P_92,0.05176431096055968
+P_86,0.7091620551640344
+P_9,0.8156592620765963
+P_90,0.4152355895738249
diff --git a/clinical_model/valid_ids.csv b/clinical_model/valid_ids.csv
new file mode 100644
index 0000000..b4e05b1
--- /dev/null
+++ b/clinical_model/valid_ids.csv
@@ -0,0 +1,105 @@
+P_384
+P_417
+P_551
+P_479
+P_555
+P_376
+P_1_26
+P_379
+P_441
+P_450
+P_760
+P_756
+P_435
+P_442
+P_766
+P_695
+P_537
+P_687
+P_467
+P_1_126
+P_1_75
+P_653
+P_1_149
+P_411
+P_358
+P_740
+P_1_81
+P_634
+P_616
+P_382
+P_342
+P_519
+P_476
+P_639
+P_759
+P_350
+P_484
+P_635
+P_507
+P_439
+P_798
+P_598
+P_414
+P_583
+P_590
+P_485
+P_511
+P_426
+P_304
+P_842
+P_797
+P_837
+P_314
+P_572
+P_641
+P_803
+P_1_55
+P_702
+P_691
+P_518
+P_1_95
+P_521
+P_589
+P_840
+P_1_131
+P_772
+P_368
+P_292
+P_750
+P_274
+P_1_146
+P_650
+P_505
+P_733
+P_623
+P_697
+P_620
+P_1_140
+P_609
+P_808
+P_526
+P_383
+P_595
+P_587
+P_580
+P_502
+P_1_22
+P_324
+P_1_72
+P_725
+P_336
+P_491
+P_781
+P_407
+P_546
+P_645
+P_425
+P_735
+P_445
+P_812
+P_385
+P_1_66
+P_743
+P_560
+P_633