baselines.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "fa7e35dd",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import itertools\n",
    "from sklearn.metrics import precision_recall_fscore_support\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from tqdm import tqdm"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2727b1ed",
   "metadata": {},
   "source": [
    "### AKI"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e8c0c4fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "from blood_panel_data_preprocessing import blood_panel_data, aki_data, sepsis_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "04aa6f67",
   "metadata": {},
   "outputs": [],
   "source": [
    "index, data, block, cost = aki_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "9937dadd",
   "metadata": {},
   "outputs": [],
   "source": [
    "from data_loader import Data_Loader\n",
    "data_loader = Data_Loader(data, block, test_ratio = 0.2, val_ratio = 0.2, index= index)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a259f823",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = data_loader.train[:,:-1]\n",
    "X_test = data_loader.test[:,:-1]\n",
    "X_val = data_loader.val[:,:-1]\n",
    "\n",
    "y_train = data_loader.train[:,-1]\n",
    "y_test = data_loader.test[:,-1]\n",
    "y_val = data_loader.val[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "9848062c",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 66/66 [11:11<00:00, 10.18s/it]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "n_estimators =  np.linspace(20,620,11,dtype=int)\n",
    "max_depths = [3,7,11,21,31,61]\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(n_estimators, max_depth))):\n",
    "    clf = RandomForestClassifier(class_weight='balanced',n_estimators = hp1, max_depth = hp2 )\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/RFAKI.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "2918c522",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>80.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.504432</td>\n",
       "      <td>0.785962</td>\n",
       "      <td>0.439320</td>\n",
       "      <td>0.763929</td>\n",
       "      <td>0.418014</td>\n",
       "      <td>0.462916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>20.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.500406</td>\n",
       "      <td>0.781915</td>\n",
       "      <td>0.416157</td>\n",
       "      <td>0.756495</td>\n",
       "      <td>0.399061</td>\n",
       "      <td>0.434783</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>260.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.500402</td>\n",
       "      <td>0.792163</td>\n",
       "      <td>0.454321</td>\n",
       "      <td>0.775117</td>\n",
       "      <td>0.439141</td>\n",
       "      <td>0.470588</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>380.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.499345</td>\n",
       "      <td>0.790910</td>\n",
       "      <td>0.452611</td>\n",
       "      <td>0.770829</td>\n",
       "      <td>0.363919</td>\n",
       "      <td>0.598465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>140.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.499343</td>\n",
       "      <td>0.789083</td>\n",
       "      <td>0.453307</td>\n",
       "      <td>0.770194</td>\n",
       "      <td>0.365777</td>\n",
       "      <td>0.595908</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1   hp2    val_f1   val_auc   test_f1  test_auc  test_prec  \\\n",
       "8    80.0  11.0  0.504432  0.785962  0.439320  0.763929   0.418014   \n",
       "2    20.0  11.0  0.500406  0.781915  0.416157  0.756495   0.399061   \n",
       "26  260.0  11.0  0.500402  0.792163  0.454321  0.775117   0.439141   \n",
       "37  380.0   7.0  0.499345  0.790910  0.452611  0.770829   0.363919   \n",
       "13  140.0   7.0  0.499343  0.789083  0.453307  0.770194   0.365777   \n",
       "\n",
       "    test_recall  \n",
       "8      0.462916  \n",
       "2      0.434783  \n",
       "26     0.470588  \n",
       "37     0.598465  \n",
       "13     0.595908  "
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fe0c0413",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "C =  [0.0001,0.001,0.1,1,10,100]\n",
    "penalty = ['l1', 'l2', 'elasticnet', 'none']\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(C, penalty))):\n",
    "    clf = LogisticRegression(class_weight='balanced',C = hp1, penalty = hp2, solver='saga', max_iter=1000, l1_ratio=0.5)\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/LRAKI.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "8bf3b2ae",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.474725</td>\n",
       "      <td>0.798561</td>\n",
       "      <td>0.452303</td>\n",
       "      <td>0.797081</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.703325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.1</td>\n",
       "      <td>none</td>\n",
       "      <td>0.474725</td>\n",
       "      <td>0.798619</td>\n",
       "      <td>0.452675</td>\n",
       "      <td>0.797155</td>\n",
       "      <td>0.333738</td>\n",
       "      <td>0.703325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>100.0</td>\n",
       "      <td>elasticnet</td>\n",
       "      <td>0.474725</td>\n",
       "      <td>0.798620</td>\n",
       "      <td>0.452675</td>\n",
       "      <td>0.797153</td>\n",
       "      <td>0.333738</td>\n",
       "      <td>0.703325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>100.0</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.474725</td>\n",
       "      <td>0.798623</td>\n",
       "      <td>0.452675</td>\n",
       "      <td>0.797157</td>\n",
       "      <td>0.333738</td>\n",
       "      <td>0.703325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>100.0</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.474725</td>\n",
       "      <td>0.798618</td>\n",
       "      <td>0.452675</td>\n",
       "      <td>0.797149</td>\n",
       "      <td>0.333738</td>\n",
       "      <td>0.703325</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1         hp2    val_f1   val_auc   test_f1  test_auc  test_prec  \\\n",
       "12    1.0          l1  0.474725  0.798561  0.452303  0.797081   0.333333   \n",
       "11    0.1        none  0.474725  0.798619  0.452675  0.797155   0.333738   \n",
       "22  100.0  elasticnet  0.474725  0.798620  0.452675  0.797153   0.333738   \n",
       "21  100.0          l2  0.474725  0.798623  0.452675  0.797157   0.333738   \n",
       "20  100.0          l1  0.474725  0.798618  0.452675  0.797149   0.333738   \n",
       "\n",
       "    test_recall  \n",
       "12     0.703325  \n",
       "11     0.703325  \n",
       "22     0.703325  \n",
       "21     0.703325  \n",
       "20     0.703325  "
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "9440eb36",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 66/66 [05:11<00:00,  4.72s/it]\n"
     ]
    }
   ],
   "source": [
    "from xgboost import XGBClassifier\n",
    "n_estimators =  np.linspace(20,620,11,dtype=int)\n",
    "max_depths = [3,7,11,21,31,61]\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(n_estimators, max_depth))):\n",
    "    clf = XGBClassifier(n_estimators = hp1, max_depth = hp2 )\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/XGAKI.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "548f1b2c",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>440.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.440567</td>\n",
       "      <td>0.786340</td>\n",
       "      <td>0.404605</td>\n",
       "      <td>0.784796</td>\n",
       "      <td>0.566820</td>\n",
       "      <td>0.314578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>500.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.436681</td>\n",
       "      <td>0.784338</td>\n",
       "      <td>0.411862</td>\n",
       "      <td>0.783264</td>\n",
       "      <td>0.578704</td>\n",
       "      <td>0.319693</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>380.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.436205</td>\n",
       "      <td>0.787949</td>\n",
       "      <td>0.422259</td>\n",
       "      <td>0.784493</td>\n",
       "      <td>0.586364</td>\n",
       "      <td>0.329923</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>320.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.428256</td>\n",
       "      <td>0.788449</td>\n",
       "      <td>0.413223</td>\n",
       "      <td>0.788339</td>\n",
       "      <td>0.584112</td>\n",
       "      <td>0.319693</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>80.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.426667</td>\n",
       "      <td>0.777801</td>\n",
       "      <td>0.391003</td>\n",
       "      <td>0.766269</td>\n",
       "      <td>0.604278</td>\n",
       "      <td>0.289003</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1  hp2    val_f1   val_auc   test_f1  test_auc  test_prec  test_recall\n",
       "42  440.0  3.0  0.440567  0.786340  0.404605  0.784796   0.566820     0.314578\n",
       "48  500.0  3.0  0.436681  0.784338  0.411862  0.783264   0.578704     0.319693\n",
       "36  380.0  3.0  0.436205  0.787949  0.422259  0.784493   0.586364     0.329923\n",
       "30  320.0  3.0  0.428256  0.788449  0.413223  0.788339   0.584112     0.319693\n",
       "7    80.0  7.0  0.426667  0.777801  0.391003  0.766269   0.604278     0.289003"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "ab0cfcd8",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 66/66 [00:57<00:00,  1.16it/s]\n"
     ]
    }
   ],
   "source": [
    "from lightgbm import LGBMClassifier\n",
    "n_estimators =  np.linspace(20,620,11,dtype=int)\n",
    "max_depths = [3,7,11,21,31,61]\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(n_estimators, max_depth))):\n",
    "    clf = LGBMClassifier(n_estimators = hp1, max_depth = hp2, class_weight= 'balanced' )\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/LGAKI.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "78ac0425",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>380.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>0.523577</td>\n",
       "      <td>0.790138</td>\n",
       "      <td>0.474453</td>\n",
       "      <td>0.789645</td>\n",
       "      <td>0.452436</td>\n",
       "      <td>0.498721</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>380.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>0.523577</td>\n",
       "      <td>0.790138</td>\n",
       "      <td>0.474453</td>\n",
       "      <td>0.789645</td>\n",
       "      <td>0.452436</td>\n",
       "      <td>0.498721</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>380.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>0.523577</td>\n",
       "      <td>0.790138</td>\n",
       "      <td>0.474453</td>\n",
       "      <td>0.789645</td>\n",
       "      <td>0.452436</td>\n",
       "      <td>0.498721</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>380.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.523002</td>\n",
       "      <td>0.794636</td>\n",
       "      <td>0.472906</td>\n",
       "      <td>0.787020</td>\n",
       "      <td>0.456057</td>\n",
       "      <td>0.491049</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>440.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.522314</td>\n",
       "      <td>0.791650</td>\n",
       "      <td>0.465823</td>\n",
       "      <td>0.783818</td>\n",
       "      <td>0.461153</td>\n",
       "      <td>0.470588</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1   hp2    val_f1   val_auc   test_f1  test_auc  test_prec  \\\n",
       "41  380.0  61.0  0.523577  0.790138  0.474453  0.789645   0.452436   \n",
       "40  380.0  31.0  0.523577  0.790138  0.474453  0.789645   0.452436   \n",
       "39  380.0  21.0  0.523577  0.790138  0.474453  0.789645   0.452436   \n",
       "38  380.0  11.0  0.523002  0.794636  0.472906  0.787020   0.456057   \n",
       "44  440.0  11.0  0.522314  0.791650  0.465823  0.783818   0.461153   \n",
       "\n",
       "    test_recall  \n",
       "41     0.498721  \n",
       "40     0.498721  \n",
       "39     0.498721  \n",
       "38     0.491049  \n",
       "44     0.470588  "
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0e96776b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "65c91d20",
   "metadata": {},
   "source": [
    "### Sepsis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "60515576",
   "metadata": {},
   "outputs": [],
   "source": [
    "from blood_panel_data_preprocessing import blood_panel_data, aki_data, sepsis_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "b0015e84",
   "metadata": {},
   "outputs": [],
   "source": [
    "data, block, cost = sepsis_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "6d6f220f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from data_loader import Data_Loader\n",
    "data_loader = Data_Loader(data, block, test_ratio = 0.2, val_ratio = 0.2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "bc5d430b",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = data_loader.train[:,:-1]\n",
    "X_test = data_loader.test[:,:-1]\n",
    "X_val = data_loader.val[:,:-1]\n",
    "\n",
    "y_train = data_loader.train[:,-1]\n",
    "y_test = data_loader.test[:,-1]\n",
    "y_val = data_loader.val[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "0ae7c792",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 66/66 [02:39<00:00,  2.42s/it]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "n_estimators =  np.linspace(20,620,11,dtype=int)\n",
    "max_depths = [3,7,11,21,31,61]\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(n_estimators, max_depth))):\n",
    "    clf = RandomForestClassifier(class_weight='balanced',n_estimators = hp1, max_depth = hp2 )\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/RFSepsis.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "f9dd4a55",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>80.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.437908</td>\n",
       "      <td>0.766100</td>\n",
       "      <td>0.456140</td>\n",
       "      <td>0.800963</td>\n",
       "      <td>0.374101</td>\n",
       "      <td>0.584270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>200.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.422442</td>\n",
       "      <td>0.759139</td>\n",
       "      <td>0.480349</td>\n",
       "      <td>0.809310</td>\n",
       "      <td>0.392857</td>\n",
       "      <td>0.617978</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>620.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.421525</td>\n",
       "      <td>0.788380</td>\n",
       "      <td>0.423077</td>\n",
       "      <td>0.830291</td>\n",
       "      <td>0.492537</td>\n",
       "      <td>0.370787</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>320.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.416667</td>\n",
       "      <td>0.790831</td>\n",
       "      <td>0.389610</td>\n",
       "      <td>0.830887</td>\n",
       "      <td>0.461538</td>\n",
       "      <td>0.337079</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>380.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.410774</td>\n",
       "      <td>0.765872</td>\n",
       "      <td>0.477477</td>\n",
       "      <td>0.806856</td>\n",
       "      <td>0.398496</td>\n",
       "      <td>0.595506</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1  hp2    val_f1   val_auc   test_f1  test_auc  test_prec  test_recall\n",
       "6    80.0  3.0  0.437908  0.766100  0.456140  0.800963   0.374101     0.584270\n",
       "18  200.0  3.0  0.422442  0.759139  0.480349  0.809310   0.392857     0.617978\n",
       "61  620.0  7.0  0.421525  0.788380  0.423077  0.830291   0.492537     0.370787\n",
       "31  320.0  7.0  0.416667  0.790831  0.389610  0.830887   0.461538     0.337079\n",
       "36  380.0  3.0  0.410774  0.765872  0.477477  0.806856   0.398496     0.595506"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "34b6f1e1",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|                                                    | 0/24 [00:00<?, ?it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "  8%|███▋                                        | 2/24 [00:00<00:01, 18.01it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1322: UserWarning: Setting penalty='none' will ignore the C and l1_ratio parameters\n",
      "  warnings.warn(\n",
      " 17%|███████▎                                    | 4/24 [00:00<00:01, 17.37it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1248: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 due to no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 25%|███████████                                 | 6/24 [00:00<00:00, 18.00it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1322: UserWarning: Setting penalty='none' will ignore the C and l1_ratio parameters\n",
      "  warnings.warn(\n",
      " 33%|██████████████▋                             | 8/24 [00:00<00:01, 15.63it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 42%|█████████████████▉                         | 10/24 [00:00<00:00, 14.29it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1322: UserWarning: Setting penalty='none' will ignore the C and l1_ratio parameters\n",
      "  warnings.warn(\n",
      " 50%|█████████████████████▌                     | 12/24 [00:00<00:00, 12.80it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 58%|█████████████████████████                  | 14/24 [00:01<00:00, 12.66it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 67%|████████████████████████████▋              | 16/24 [00:01<00:00, 12.46it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 75%|████████████████████████████████▎          | 18/24 [00:01<00:00, 12.37it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1322: UserWarning: Setting penalty='none' will ignore the C and l1_ratio parameters\n",
      "  warnings.warn(\n",
      " 83%|███████████████████████████████████▊       | 20/24 [00:01<00:00, 11.85it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 92%|███████████████████████████████████████▍   | 22/24 [00:01<00:00, 11.80it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1322: UserWarning: Setting penalty='none' will ignore the C and l1_ratio parameters\n",
      "  warnings.warn(\n",
      "100%|███████████████████████████████████████████| 24/24 [00:01<00:00, 12.64it/s]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "C =  [0.0001,0.001,0.1,1,10,100]\n",
    "penalty = ['l1', 'l2', 'elasticnet', 'none']\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(C, penalty))):\n",
    "    clf = LogisticRegression(class_weight='balanced',C = hp1, penalty = hp2, solver='saga', max_iter=1000, l1_ratio=0.5)\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/LRSepsis.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "aed9d96d",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.1</td>\n",
       "      <td>elasticnet</td>\n",
       "      <td>0.455882</td>\n",
       "      <td>0.790852</td>\n",
       "      <td>0.505929</td>\n",
       "      <td>0.824994</td>\n",
       "      <td>0.390244</td>\n",
       "      <td>0.719101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.455206</td>\n",
       "      <td>0.791390</td>\n",
       "      <td>0.505929</td>\n",
       "      <td>0.823848</td>\n",
       "      <td>0.390244</td>\n",
       "      <td>0.719101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.455206</td>\n",
       "      <td>0.791566</td>\n",
       "      <td>0.498024</td>\n",
       "      <td>0.823137</td>\n",
       "      <td>0.384146</td>\n",
       "      <td>0.707865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>100.0</td>\n",
       "      <td>elasticnet</td>\n",
       "      <td>0.455206</td>\n",
       "      <td>0.791514</td>\n",
       "      <td>0.498024</td>\n",
       "      <td>0.822885</td>\n",
       "      <td>0.384146</td>\n",
       "      <td>0.707865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>100.0</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.455206</td>\n",
       "      <td>0.791493</td>\n",
       "      <td>0.498024</td>\n",
       "      <td>0.822908</td>\n",
       "      <td>0.384146</td>\n",
       "      <td>0.707865</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1         hp2    val_f1   val_auc   test_f1  test_auc  test_prec  \\\n",
       "10    0.1  elasticnet  0.455882  0.790852  0.505929  0.824994   0.390244   \n",
       "12    1.0          l1  0.455206  0.791390  0.505929  0.823848   0.390244   \n",
       "13    1.0          l2  0.455206  0.791566  0.498024  0.823137   0.384146   \n",
       "22  100.0  elasticnet  0.455206  0.791514  0.498024  0.822885   0.384146   \n",
       "21  100.0          l2  0.455206  0.791493  0.498024  0.822908   0.384146   \n",
       "\n",
       "    test_recall  \n",
       "10     0.719101  \n",
       "12     0.719101  \n",
       "13     0.707865  \n",
       "22     0.707865  \n",
       "21     0.707865  "
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "680bd196",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 66/66 [01:45<00:00,  1.60s/it]\n"
     ]
    }
   ],
   "source": [
    "from xgboost import XGBClassifier\n",
    "n_estimators =  np.linspace(20,620,11,dtype=int)\n",
    "max_depths = [3,7,11,21,31,61]\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(n_estimators, max_depth))):\n",
    "    clf = XGBClassifier(n_estimators = hp1, max_depth = hp2 )\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/XGSepsis.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "88a9a120",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>260.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.373737</td>\n",
       "      <td>0.781636</td>\n",
       "      <td>0.430769</td>\n",
       "      <td>0.827540</td>\n",
       "      <td>0.682927</td>\n",
       "      <td>0.314607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>140.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.369231</td>\n",
       "      <td>0.790904</td>\n",
       "      <td>0.459259</td>\n",
       "      <td>0.829466</td>\n",
       "      <td>0.673913</td>\n",
       "      <td>0.348315</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>80.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.366492</td>\n",
       "      <td>0.799561</td>\n",
       "      <td>0.390625</td>\n",
       "      <td>0.832653</td>\n",
       "      <td>0.641026</td>\n",
       "      <td>0.280899</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>200.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.336842</td>\n",
       "      <td>0.783022</td>\n",
       "      <td>0.466165</td>\n",
       "      <td>0.825843</td>\n",
       "      <td>0.704545</td>\n",
       "      <td>0.348315</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>620.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.336634</td>\n",
       "      <td>0.770217</td>\n",
       "      <td>0.415385</td>\n",
       "      <td>0.825086</td>\n",
       "      <td>0.658537</td>\n",
       "      <td>0.303371</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1  hp2    val_f1   val_auc   test_f1  test_auc  test_prec  test_recall\n",
       "24  260.0  3.0  0.373737  0.781636  0.430769  0.827540   0.682927     0.314607\n",
       "12  140.0  3.0  0.369231  0.790904  0.459259  0.829466   0.673913     0.348315\n",
       "6    80.0  3.0  0.366492  0.799561  0.390625  0.832653   0.641026     0.280899\n",
       "18  200.0  3.0  0.336842  0.783022  0.466165  0.825843   0.704545     0.348315\n",
       "60  620.0  3.0  0.336634  0.770217  0.415385  0.825086   0.658537     0.303371"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "47d27ce2",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 66/66 [00:41<00:00,  1.59it/s]\n"
     ]
    }
   ],
   "source": [
    "from lightgbm import LGBMClassifier\n",
    "n_estimators =  np.linspace(20,620,11,dtype=int)\n",
    "max_depths = [3,7,11,21,31,61]\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(n_estimators, max_depth))):\n",
    "    clf = LGBMClassifier(n_estimators = hp1, max_depth = hp2, class_weight= 'balanced' )\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/LGSepsis.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "652aae01",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>80.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.464183</td>\n",
       "      <td>0.778523</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.843729</td>\n",
       "      <td>0.410072</td>\n",
       "      <td>0.640449</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>200.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.463415</td>\n",
       "      <td>0.778957</td>\n",
       "      <td>0.502415</td>\n",
       "      <td>0.846549</td>\n",
       "      <td>0.440678</td>\n",
       "      <td>0.584270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>140.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.461538</td>\n",
       "      <td>0.780695</td>\n",
       "      <td>0.497696</td>\n",
       "      <td>0.846595</td>\n",
       "      <td>0.421875</td>\n",
       "      <td>0.606742</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>260.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.461538</td>\n",
       "      <td>0.776723</td>\n",
       "      <td>0.525253</td>\n",
       "      <td>0.842055</td>\n",
       "      <td>0.477064</td>\n",
       "      <td>0.584270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>320.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.441558</td>\n",
       "      <td>0.775792</td>\n",
       "      <td>0.529101</td>\n",
       "      <td>0.841023</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.561798</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1  hp2    val_f1   val_auc   test_f1  test_auc  test_prec  test_recall\n",
       "6    80.0  3.0  0.464183  0.778523  0.500000  0.843729   0.410072     0.640449\n",
       "18  200.0  3.0  0.463415  0.778957  0.502415  0.846549   0.440678     0.584270\n",
       "12  140.0  3.0  0.461538  0.780695  0.497696  0.846595   0.421875     0.606742\n",
       "24  260.0  3.0  0.461538  0.776723  0.525253  0.842055   0.477064     0.584270\n",
       "30  320.0  3.0  0.441558  0.775792  0.529101  0.841023   0.500000     0.561798"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e961dcd",
   "metadata": {},
   "source": [
    "### BP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "d9788d96",
   "metadata": {},
   "outputs": [],
   "source": [
    "data, block, cost = blood_panel_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "8d1592d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "from data_loader import Data_Loader\n",
    "data_loader = Data_Loader(data, block, test_ratio = 0.2, val_ratio = 0.2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "e0bbf8e0",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = data_loader.train[:,:-1]\n",
    "X_test = data_loader.test[:,:-1]\n",
    "X_val = data_loader.val[:,:-1]\n",
    "\n",
    "y_train = data_loader.train[:,-1]\n",
    "y_test = data_loader.test[:,-1]\n",
    "y_val = data_loader.val[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "2070e5c9",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 66/66 [23:54<00:00, 21.74s/it]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "n_estimators =  np.linspace(20,620,11,dtype=int)\n",
    "max_depths = [3,7,11,21,31,61]\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(n_estimators, max_depth))):\n",
    "    clf = RandomForestClassifier(class_weight='balanced',n_estimators = hp1, max_depth = hp2 )\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/RFBP.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "5bd13834",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>320.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.649046</td>\n",
       "      <td>0.945845</td>\n",
       "      <td>0.595059</td>\n",
       "      <td>0.937782</td>\n",
       "      <td>0.515829</td>\n",
       "      <td>0.703046</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>140.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.646628</td>\n",
       "      <td>0.946550</td>\n",
       "      <td>0.604502</td>\n",
       "      <td>0.938462</td>\n",
       "      <td>0.523191</td>\n",
       "      <td>0.715736</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>500.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.644543</td>\n",
       "      <td>0.946376</td>\n",
       "      <td>0.596567</td>\n",
       "      <td>0.937597</td>\n",
       "      <td>0.516729</td>\n",
       "      <td>0.705584</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>380.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.641537</td>\n",
       "      <td>0.945990</td>\n",
       "      <td>0.593383</td>\n",
       "      <td>0.937345</td>\n",
       "      <td>0.511971</td>\n",
       "      <td>0.705584</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>200.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.640408</td>\n",
       "      <td>0.945425</td>\n",
       "      <td>0.603854</td>\n",
       "      <td>0.937120</td>\n",
       "      <td>0.522222</td>\n",
       "      <td>0.715736</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1   hp2    val_f1   val_auc   test_f1  test_auc  test_prec  \\\n",
       "32  320.0  11.0  0.649046  0.945845  0.595059  0.937782   0.515829   \n",
       "14  140.0  11.0  0.646628  0.946550  0.604502  0.938462   0.523191   \n",
       "50  500.0  11.0  0.644543  0.946376  0.596567  0.937597   0.516729   \n",
       "38  380.0  11.0  0.641537  0.945990  0.593383  0.937345   0.511971   \n",
       "20  200.0  11.0  0.640408  0.945425  0.603854  0.937120   0.522222   \n",
       "\n",
       "    test_recall  \n",
       "32     0.703046  \n",
       "14     0.715736  \n",
       "50     0.705584  \n",
       "38     0.705584  \n",
       "20     0.715736  "
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "379021ff",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|                                                    | 0/24 [00:00<?, ?it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "  4%|█▊                                          | 1/24 [00:00<00:16,  1.37it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 12%|█████▌                                      | 3/24 [00:01<00:13,  1.59it/s]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1322: UserWarning: Setting penalty='none' will ignore the C and l1_ratio parameters\n",
      "  warnings.warn(\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n",
      " 17%|███████▎                                    | 4/24 [00:25<02:31,  7.59s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 21%|█████████▏                                  | 5/24 [00:26<01:43,  5.47s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 29%|████████████▊                               | 7/24 [00:26<00:49,  2.90s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1322: UserWarning: Setting penalty='none' will ignore the C and l1_ratio parameters\n",
      "  warnings.warn(\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n",
      " 33%|██████████████▋                             | 8/24 [00:50<02:26,  9.18s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 38%|████████████████▌                           | 9/24 [00:54<01:52,  7.49s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 46%|███████████████████▋                       | 11/24 [01:00<01:08,  5.27s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1322: UserWarning: Setting penalty='none' will ignore the C and l1_ratio parameters\n",
      "  warnings.warn(\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n",
      " 50%|█████████████████████▌                     | 12/24 [01:24<02:10, 10.87s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 54%|███████████████████████▎                   | 13/24 [01:39<02:13, 12.17s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 62%|██████████████████████████▉                | 15/24 [02:03<01:47, 12.00s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n",
      " 67%|████████████████████████████▋              | 16/24 [02:27<02:05, 15.63s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n",
      " 71%|██████████████████████████████▍            | 17/24 [02:53<02:11, 18.80s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      " 75%|████████████████████████████████▎          | 18/24 [03:15<01:58, 19.79s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n",
      " 79%|██████████████████████████████████         | 19/24 [03:42<01:48, 21.80s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1322: UserWarning: Setting penalty='none' will ignore the C and l1_ratio parameters\n",
      "  warnings.warn(\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n",
      " 83%|███████████████████████████████████▊       | 20/24 [04:06<01:29, 22.40s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l1)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 88%|█████████████████████████████████████▋     | 21/24 [04:32<01:11, 23.67s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=l2)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n",
      " 92%|███████████████████████████████████████▍   | 22/24 [04:56<00:47, 23.72s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n",
      " 96%|█████████████████████████████████████████▏ | 23/24 [05:23<00:24, 24.57s/it]/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1317: UserWarning: l1_ratio parameter is only used when penalty is 'elasticnet'. Got (penalty=none)\n",
      "  warnings.warn(\"l1_ratio parameter is only used when penalty is \"\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:1322: UserWarning: Setting penalty='none' will ignore the C and l1_ratio parameters\n",
      "  warnings.warn(\n",
      "/home/ylo7832/.conda/envs/bertenv/lib/python3.8/site-packages/sklearn/linear_model/_sag.py:328: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  warnings.warn(\"The max_iter was reached which means \"\n",
      "100%|███████████████████████████████████████████| 24/24 [05:46<00:00, 14.46s/it]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "C =  [0.0001,0.001,0.1,1,10,100]\n",
    "penalty = ['l1', 'l2', 'elasticnet', 'none']\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(C, penalty))):\n",
    "    clf = LogisticRegression(class_weight='balanced',C = hp1, penalty = hp2, solver='saga', max_iter=1000, l1_ratio=0.5)\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/LRBP.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "d6dac458",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>l1</td>\n",
       "      <td>0.561497</td>\n",
       "      <td>0.938746</td>\n",
       "      <td>0.537500</td>\n",
       "      <td>0.934131</td>\n",
       "      <td>0.388262</td>\n",
       "      <td>0.873096</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.1</td>\n",
       "      <td>elasticnet</td>\n",
       "      <td>0.561497</td>\n",
       "      <td>0.939356</td>\n",
       "      <td>0.539484</td>\n",
       "      <td>0.934505</td>\n",
       "      <td>0.389831</td>\n",
       "      <td>0.875635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1.0</td>\n",
       "      <td>elasticnet</td>\n",
       "      <td>0.561197</td>\n",
       "      <td>0.938878</td>\n",
       "      <td>0.538763</td>\n",
       "      <td>0.934058</td>\n",
       "      <td>0.389581</td>\n",
       "      <td>0.873096</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>10.0</td>\n",
       "      <td>elasticnet</td>\n",
       "      <td>0.561159</td>\n",
       "      <td>0.938319</td>\n",
       "      <td>0.535938</td>\n",
       "      <td>0.933115</td>\n",
       "      <td>0.387133</td>\n",
       "      <td>0.870558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>l2</td>\n",
       "      <td>0.560897</td>\n",
       "      <td>0.938913</td>\n",
       "      <td>0.538341</td>\n",
       "      <td>0.933983</td>\n",
       "      <td>0.389140</td>\n",
       "      <td>0.873096</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     hp1         hp2    val_f1   val_auc   test_f1  test_auc  test_prec  \\\n",
       "12   1.0          l1  0.561497  0.938746  0.537500  0.934131   0.388262   \n",
       "10   0.1  elasticnet  0.561497  0.939356  0.539484  0.934505   0.389831   \n",
       "14   1.0  elasticnet  0.561197  0.938878  0.538763  0.934058   0.389581   \n",
       "18  10.0  elasticnet  0.561159  0.938319  0.535938  0.933115   0.387133   \n",
       "13   1.0          l2  0.560897  0.938913  0.538341  0.933983   0.389140   \n",
       "\n",
       "    test_recall  \n",
       "12     0.873096  \n",
       "10     0.875635  \n",
       "14     0.873096  \n",
       "18     0.870558  \n",
       "13     0.873096  "
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "9c28769c",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 66/66 [06:47<00:00,  6.18s/it]\n"
     ]
    }
   ],
   "source": [
    "from xgboost import XGBClassifier\n",
    "n_estimators =  np.linspace(20,620,11,dtype=int)\n",
    "max_depths = [3,7,11,21,31,61]\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(n_estimators, max_depth))):\n",
    "    clf = XGBClassifier(n_estimators = hp1, max_depth = hp2 )\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/XGBP.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "2d7dd7db",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>380.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.617188</td>\n",
       "      <td>0.948749</td>\n",
       "      <td>0.617143</td>\n",
       "      <td>0.938080</td>\n",
       "      <td>0.705882</td>\n",
       "      <td>0.548223</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>380.0</td>\n",
       "      <td>61.0</td>\n",
       "      <td>0.616290</td>\n",
       "      <td>0.950904</td>\n",
       "      <td>0.620290</td>\n",
       "      <td>0.944410</td>\n",
       "      <td>0.722973</td>\n",
       "      <td>0.543147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>380.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>0.616290</td>\n",
       "      <td>0.950904</td>\n",
       "      <td>0.620290</td>\n",
       "      <td>0.944410</td>\n",
       "      <td>0.722973</td>\n",
       "      <td>0.543147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>500.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.614786</td>\n",
       "      <td>0.948335</td>\n",
       "      <td>0.606581</td>\n",
       "      <td>0.938363</td>\n",
       "      <td>0.695082</td>\n",
       "      <td>0.538071</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>620.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.613437</td>\n",
       "      <td>0.947937</td>\n",
       "      <td>0.605452</td>\n",
       "      <td>0.938191</td>\n",
       "      <td>0.696370</td>\n",
       "      <td>0.535533</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1   hp2    val_f1   val_auc   test_f1  test_auc  test_prec  \\\n",
       "37  380.0   7.0  0.617188  0.948749  0.617143  0.938080   0.705882   \n",
       "41  380.0  61.0  0.616290  0.950904  0.620290  0.944410   0.722973   \n",
       "40  380.0  31.0  0.616290  0.950904  0.620290  0.944410   0.722973   \n",
       "49  500.0   7.0  0.614786  0.948335  0.606581  0.938363   0.695082   \n",
       "61  620.0   7.0  0.613437  0.947937  0.605452  0.938191   0.696370   \n",
       "\n",
       "    test_recall  \n",
       "37     0.548223  \n",
       "41     0.543147  \n",
       "40     0.543147  \n",
       "49     0.538071  \n",
       "61     0.535533  "
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "98edb1a5",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 66/66 [01:17<00:00,  1.17s/it]\n"
     ]
    }
   ],
   "source": [
    "from lightgbm import LGBMClassifier\n",
    "n_estimators =  np.linspace(20,620,11,dtype=int)\n",
    "max_depths = [3,7,11,21,31,61]\n",
    "results = pd.DataFrame(columns = ['hp1','hp2', 'val_f1', 'val_auc', 'test_f1','test_auc', 'test_prec', 'test_recall'])\n",
    "i = 0\n",
    "for hp1, hp2 in tqdm(list(itertools.product(n_estimators, max_depth))):\n",
    "    clf = LGBMClassifier(n_estimators = hp1, max_depth = hp2, class_weight= 'balanced' )\n",
    "    clf.fit(X_train, y_train)\n",
    "    y_hat_val = clf.predict_proba(X_val)[:,1]\n",
    "    y_hat_test = clf.predict_proba(X_test)[:,1]\n",
    "    results.loc[i] = [hp1, hp2,\n",
    "                       precision_recall_fscore_support(y_val, y_hat_val > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_val, y_hat_val),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[2],\n",
    "                       roc_auc_score(y_test, y_hat_test),\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[0],\n",
    "                       precision_recall_fscore_support(y_test, y_hat_test > 0.5,average='binary')[1]\n",
    "                      ]\n",
    "    i+=1\n",
    "results.to_csv('/home/ylo7832/Blood_Panel-master/baseline/LGBP.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "b46e8b45",
   "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>hp1</th>\n",
       "      <th>hp2</th>\n",
       "      <th>val_f1</th>\n",
       "      <th>val_auc</th>\n",
       "      <th>test_f1</th>\n",
       "      <th>test_auc</th>\n",
       "      <th>test_prec</th>\n",
       "      <th>test_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>380.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.675266</td>\n",
       "      <td>0.952017</td>\n",
       "      <td>0.626932</td>\n",
       "      <td>0.940576</td>\n",
       "      <td>0.554688</td>\n",
       "      <td>0.720812</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>320.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.672071</td>\n",
       "      <td>0.947787</td>\n",
       "      <td>0.627282</td>\n",
       "      <td>0.943151</td>\n",
       "      <td>0.543762</td>\n",
       "      <td>0.741117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>620.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.671010</td>\n",
       "      <td>0.947281</td>\n",
       "      <td>0.636472</td>\n",
       "      <td>0.942239</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.677665</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>560.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.669348</td>\n",
       "      <td>0.952441</td>\n",
       "      <td>0.631829</td>\n",
       "      <td>0.942014</td>\n",
       "      <td>0.593750</td>\n",
       "      <td>0.675127</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>320.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.669118</td>\n",
       "      <td>0.951881</td>\n",
       "      <td>0.625806</td>\n",
       "      <td>0.941327</td>\n",
       "      <td>0.542910</td>\n",
       "      <td>0.738579</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      hp1   hp2    val_f1   val_auc   test_f1  test_auc  test_prec  \\\n",
       "38  380.0  11.0  0.675266  0.952017  0.626932  0.940576   0.554688   \n",
       "31  320.0   7.0  0.672071  0.947787  0.627282  0.943151   0.543762   \n",
       "61  620.0   7.0  0.671010  0.947281  0.636472  0.942239   0.600000   \n",
       "56  560.0  11.0  0.669348  0.952441  0.631829  0.942014   0.593750   \n",
       "32  320.0  11.0  0.669118  0.951881  0.625806  0.941327   0.542910   \n",
       "\n",
       "    test_recall  \n",
       "38     0.720812  \n",
       "31     0.741117  \n",
       "61     0.677665  \n",
       "56     0.675127  \n",
       "32     0.738579  "
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.sort_values('val_f1', ascending=False).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0c456ea2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "233e212f",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "bertenv",
   "language": "python",
   "name": "bertenv"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}