model_tuning.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "7a628b47",
   "metadata": {},
   "source": [
    "# Datathon"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "119f6c66",
   "metadata": {},
   "source": [
    "*import libraries*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "28128733",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import libraries here\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "import seaborn as sns\n",
    "pd.options.mode.chained_assignment = None  # default='warn'\n",
    "\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV\n",
    "from sklearn.preprocessing import StandardScaler, PolynomialFeatures\n",
    "from sklearn.linear_model import (RidgeCV, \n",
    "                                  LassoCV, \n",
    "                                  ElasticNetCV, \n",
    "                                  LogisticRegressionCV,\n",
    "                                  LinearRegression,\n",
    "                                  LogisticRegression)\n",
    "from imblearn.over_sampling import SMOTE\n",
    "from imblearn.pipeline import Pipeline\n",
    "from sklearn.metrics import (confusion_matrix, \n",
    "                             plot_confusion_matrix, \n",
    "                             roc_auc_score, \n",
    "                             plot_roc_curve, \n",
    "                             accuracy_score,\n",
    "                            )\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier, ExtraTreesClassifier, BaggingClassifier, AdaBoostClassifier\n",
    "from sklearn.svm import LinearSVC, SVC\n",
    "\n",
    "import xgboost as xgb\n",
    "import time\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "70811b5a",
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.max_rows', 100)\n",
    "pd.set_option('display.max_columns', 50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "aeb7583f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import data\n",
    "df = pd.read_csv('./data//final_df.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "07a8ee8a",
   "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>Unnamed: 0</th>\n",
       "      <th>subject_id</th>\n",
       "      <th>gender</th>\n",
       "      <th>mortality_deathtime</th>\n",
       "      <th>anchor_age</th>\n",
       "      <th>icu_intime_dt</th>\n",
       "      <th>aspirin_timediff</th>\n",
       "      <th>p2y12_timediff</th>\n",
       "      <th>antiplatelet_timediff</th>\n",
       "      <th>hemo_difftime</th>\n",
       "      <th>hemo_valuenum</th>\n",
       "      <th>trop_risetime</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>10000980</td>\n",
       "      <td>F</td>\n",
       "      <td>0</td>\n",
       "      <td>73</td>\n",
       "      <td>2189-06-27 08:42:00</td>\n",
       "      <td>3.366667</td>\n",
       "      <td>3.366667</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2189-06-27 14:18:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>10002013</td>\n",
       "      <td>F</td>\n",
       "      <td>0</td>\n",
       "      <td>53</td>\n",
       "      <td>2160-05-18 10:00:00</td>\n",
       "      <td>2.250000</td>\n",
       "      <td>2.250000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>10002430</td>\n",
       "      <td>M</td>\n",
       "      <td>0</td>\n",
       "      <td>86</td>\n",
       "      <td>2129-06-13 00:43:00</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>10002495</td>\n",
       "      <td>M</td>\n",
       "      <td>0</td>\n",
       "      <td>81</td>\n",
       "      <td>2141-05-22 20:18:00</td>\n",
       "      <td>11.716667</td>\n",
       "      <td>9.716667</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.716667</td>\n",
       "      <td>11.1</td>\n",
       "      <td>2141-05-22 23:05:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>10002527</td>\n",
       "      <td>M</td>\n",
       "      <td>0</td>\n",
       "      <td>50</td>\n",
       "      <td>2136-03-24 10:24:00</td>\n",
       "      <td>21.500000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0  subject_id gender  mortality_deathtime  anchor_age  \\\n",
       "0           0    10000980      F                    0          73   \n",
       "1           1    10002013      F                    0          53   \n",
       "2           2    10002430      M                    0          86   \n",
       "3           3    10002495      M                    0          81   \n",
       "4           4    10002527      M                    0          50   \n",
       "\n",
       "         icu_intime_dt  aspirin_timediff  p2y12_timediff  \\\n",
       "0  2189-06-27 08:42:00          3.366667        3.366667   \n",
       "1  2160-05-18 10:00:00          2.250000        2.250000   \n",
       "2  2129-06-13 00:43:00          4.000000             NaN   \n",
       "3  2141-05-22 20:18:00         11.716667        9.716667   \n",
       "4  2136-03-24 10:24:00         21.500000             NaN   \n",
       "\n",
       "   antiplatelet_timediff  hemo_difftime  hemo_valuenum        trop_risetime  \n",
       "0                    NaN            NaN            NaN  2189-06-27 14:18:00  \n",
       "1                    NaN            NaN            NaN                  NaN  \n",
       "2                    NaN            NaN            NaN                  NaN  \n",
       "3                    NaN       0.716667           11.1  2141-05-22 23:05:00  \n",
       "4                    NaN            NaN            NaN                  NaN  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "d8cbf6dd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 16407 entries, 0 to 16406\n",
      "Data columns (total 12 columns):\n",
      " #   Column                 Non-Null Count  Dtype  \n",
      "---  ------                 --------------  -----  \n",
      " 0   Unnamed: 0             16407 non-null  int64  \n",
      " 1   subject_id             16407 non-null  int64  \n",
      " 2   gender                 16407 non-null  object \n",
      " 3   mortality_deathtime    16407 non-null  int64  \n",
      " 4   anchor_age             16407 non-null  int64  \n",
      " 5   icu_intime_dt          16407 non-null  object \n",
      " 6   aspirin_timediff       15818 non-null  float64\n",
      " 7   p2y12_timediff         3293 non-null   float64\n",
      " 8   antiplatelet_timediff  22 non-null     float64\n",
      " 9   hemo_difftime          5105 non-null   float64\n",
      " 10  hemo_valuenum          5105 non-null   float64\n",
      " 11  trop_risetime          5147 non-null   object \n",
      "dtypes: float64(5), int64(4), object(3)\n",
      "memory usage: 1.5+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "86d578f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert datetime\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "20306f3e",
   "metadata": {},
   "outputs": [],
   "source": [
    "df.rename(columns={'mortality_deathtime': 'mortality'},inplace=True) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "123053fb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['F', 'M'], dtype=object)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['gender'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d81aea09",
   "metadata": {},
   "outputs": [],
   "source": [
    "# convert gender\n",
    "df['gender_bin'] = df['gender'].apply(lambda x: 1 if x is 'M' else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "2970b667",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.929055\n",
       "1    0.070945\n",
       "Name: mortality, dtype: float64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['mortality'].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2af917a6",
   "metadata": {},
   "source": [
    "### Baseline model (with SMOTE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "cbb415b6",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_sm = df[['anchor_age']]\n",
    "y_sm = df[['mortality']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "96e71dbf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mortality\n",
       "0            0.929055\n",
       "1            0.070945\n",
       "dtype: float64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# imbalanced \n",
    "y_sm.value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "198e50d3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# split data\n",
    "Xsm_train, Xsm_test, ysm_train, ysm_test = train_test_split(X_sm, y_sm, test_size = 0.33, random_state=42, stratify=y_sm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "62e123e0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# scaling\n",
    "ss = StandardScaler()\n",
    "\n",
    "Xsm_train_sc = ss.fit_transform(Xsm_train)\n",
    "Xsm_test_sc = ss.transform(Xsm_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "b4a42c77",
   "metadata": {},
   "outputs": [],
   "source": [
    "# create synthetic data for training set\n",
    "smote = SMOTE()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "15ec3696",
   "metadata": {},
   "outputs": [],
   "source": [
    "Xsmote_train, ysmote_train = smote.fit_resample(Xsm_train_sc, ysm_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "45b8b08e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<bound method SamplerMixin.fit of SMOTE()>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "smote.fit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "346de33f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "mortality\n",
       "0            10212\n",
       "1            10212\n",
       "dtype: int64"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# balanced\n",
    "ysmote_train.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "6873513b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.6056600078339209, 0.5850415512465375)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# instantiate\n",
    "lr = LogisticRegression()\n",
    "\n",
    "# fit model\n",
    "lr.fit(Xsmote_train, ysmote_train)\n",
    "\n",
    "# score\n",
    "lr.score(Xsmote_train, ysmote_train), lr.score(Xsm_test_sc, ysm_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d170659",
   "metadata": {},
   "source": [
    "So the **baseline accuracy** is 0.61 (for train set), and 0.59 (for test set)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bf70958d",
   "metadata": {},
   "source": [
    "## Feature selection"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cbb01546",
   "metadata": {},
   "source": [
    "### Logistic regression (with SMOTE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "d8b134de",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 16407 entries, 0 to 16406\n",
      "Data columns (total 13 columns):\n",
      " #   Column                 Non-Null Count  Dtype  \n",
      "---  ------                 --------------  -----  \n",
      " 0   Unnamed: 0             16407 non-null  int64  \n",
      " 1   subject_id             16407 non-null  int64  \n",
      " 2   gender                 16407 non-null  object \n",
      " 3   mortality              16407 non-null  int64  \n",
      " 4   anchor_age             16407 non-null  int64  \n",
      " 5   icu_intime_dt          16407 non-null  object \n",
      " 6   aspirin_timediff       15818 non-null  float64\n",
      " 7   p2y12_timediff         3293 non-null   float64\n",
      " 8   antiplatelet_timediff  22 non-null     float64\n",
      " 9   hemo_difftime          5105 non-null   float64\n",
      " 10  hemo_valuenum          5105 non-null   float64\n",
      " 11  trop_risetime          5147 non-null   object \n",
      " 12  gender_bin             16407 non-null  int64  \n",
      "dtypes: float64(5), int64(5), object(3)\n",
      "memory usage: 1.6+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "d9cc15fe",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.fillna(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "c0e3da14",
   "metadata": {},
   "outputs": [],
   "source": [
    "# from previous notebook\n",
    "# drop the above columns\n",
    "df.drop(columns=['Unnamed: 0','subject_id','gender','icu_intime_dt','trop_risetime'],inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "600dd68b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(16407, 8)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "ae26b167",
   "metadata": {},
   "outputs": [],
   "source": [
    "# using (most) entire dataset \n",
    "X = df.drop(columns=['mortality'])\n",
    "y = df[['mortality']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "1af265ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "# split data\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.33, random_state = 42, stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "66bccf11",
   "metadata": {},
   "outputs": [],
   "source": [
    "pipe = Pipeline([\n",
    "        ('scale', StandardScaler()),\n",
    "        ('sampling', SMOTE(sampling_strategy='minority')),\n",
    "        ('logreg', LogisticRegression(max_iter=1_000, solver='saga'))\n",
    "    ])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "27d67592",
   "metadata": {},
   "outputs": [],
   "source": [
    "# regardless of parameters, there is overfitting.\n",
    "pipe_params = {\n",
    "    'sampling__k_neighbors': [5, 10],   # \n",
    "    'logreg__penalty': ['l2', 'l1'],   # \n",
    "    'logreg__C': [1] # \n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "0d3bfb6f",
   "metadata": {},
   "outputs": [],
   "source": [
    "grid = GridSearchCV(pipe, pipe_params, scoring='roc_auc', n_jobs=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "ca693d02",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "anchor_age               0\n",
       "aspirin_timediff         0\n",
       "p2y12_timediff           0\n",
       "antiplatelet_timediff    0\n",
       "hemo_difftime            0\n",
       "hemo_valuenum            0\n",
       "gender_bin               0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "303b9bcc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 10992 entries, 13492 to 9663\n",
      "Data columns (total 7 columns):\n",
      " #   Column                 Non-Null Count  Dtype  \n",
      "---  ------                 --------------  -----  \n",
      " 0   anchor_age             10992 non-null  int64  \n",
      " 1   aspirin_timediff       10992 non-null  float64\n",
      " 2   p2y12_timediff         10992 non-null  float64\n",
      " 3   antiplatelet_timediff  10992 non-null  float64\n",
      " 4   hemo_difftime          10992 non-null  float64\n",
      " 5   hemo_valuenum          10992 non-null  float64\n",
      " 6   gender_bin             10992 non-null  int64  \n",
      "dtypes: float64(5), int64(2)\n",
      "memory usage: 687.0 KB\n"
     ]
    }
   ],
   "source": [
    "X_train.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "5caa1603",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(estimator=Pipeline(steps=[('scale', StandardScaler()),\n",
       "                                       ('sampling',\n",
       "                                        SMOTE(sampling_strategy='minority')),\n",
       "                                       ('logreg',\n",
       "                                        LogisticRegression(max_iter=1000,\n",
       "                                                           solver='saga'))]),\n",
       "             n_jobs=-1,\n",
       "             param_grid={'logreg__C': [1], 'logreg__penalty': ['l2', 'l1'],\n",
       "                         'sampling__k_neighbors': [5, 10]},\n",
       "             scoring='roc_auc')"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "3bb13746",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'logreg__C': 1, 'logreg__penalty': 'l1', 'sampling__k_neighbors': 5}"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "60f6a151",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.7448539676800546, 0.7297624519644869)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid.score(X_train, y_train), grid.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "232dc584",
   "metadata": {},
   "source": [
    "Accuracy for train set is 0.74, while test set is 0.723. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "4bedfd57",
   "metadata": {},
   "outputs": [],
   "source": [
    "# getting feature importance \n",
    "lr_coef = grid.best_estimator_.named_steps['logreg'].coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "97a65b4a",
   "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>variable</th>\n",
       "      <th>coef</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>hemo_valuenum</td>\n",
       "      <td>0.793366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>anchor_age</td>\n",
       "      <td>0.520775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>antiplatelet_timediff</td>\n",
       "      <td>0.029590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>gender_bin</td>\n",
       "      <td>-0.027748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>p2y12_timediff</td>\n",
       "      <td>-0.033396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>aspirin_timediff</td>\n",
       "      <td>-0.061185</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>hemo_difftime</td>\n",
       "      <td>-0.111736</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                variable      coef\n",
       "5          hemo_valuenum  0.793366\n",
       "0             anchor_age  0.520775\n",
       "3  antiplatelet_timediff  0.029590\n",
       "6             gender_bin -0.027748\n",
       "2         p2y12_timediff -0.033396\n",
       "1       aspirin_timediff -0.061185\n",
       "4          hemo_difftime -0.111736"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coef_tot = pd.DataFrame(list(zip(X.columns,lr_coef[0])))\n",
    "coef_tot.columns = ['variable','coef']\n",
    "coef_tot.sort_values(by='coef',ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0ce4bb55",
   "metadata": {},
   "source": [
    "**top 5 features**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "3b93d9d4",
   "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>variable</th>\n",
       "      <th>coef</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>hemo_valuenum</td>\n",
       "      <td>0.793366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>anchor_age</td>\n",
       "      <td>0.520775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>antiplatelet_timediff</td>\n",
       "      <td>0.029590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>gender_bin</td>\n",
       "      <td>-0.027748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>p2y12_timediff</td>\n",
       "      <td>-0.033396</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                variable      coef\n",
       "5          hemo_valuenum  0.793366\n",
       "0             anchor_age  0.520775\n",
       "3  antiplatelet_timediff  0.029590\n",
       "6             gender_bin -0.027748\n",
       "2         p2y12_timediff -0.033396"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coef_tot.sort_values(by='coef',ascending=False).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0d4687a2",
   "metadata": {},
   "source": [
    "There is ?? higher likelihood to die with one unit increase in hemo_valuenum, compared to other features"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a1bbf5f0",
   "metadata": {},
   "source": [
    "**negative features**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "466220ee",
   "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>variable</th>\n",
       "      <th>coef</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>hemo_difftime</td>\n",
       "      <td>-0.111736</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>aspirin_timediff</td>\n",
       "      <td>-0.061185</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>p2y12_timediff</td>\n",
       "      <td>-0.033396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>gender_bin</td>\n",
       "      <td>-0.027748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>antiplatelet_timediff</td>\n",
       "      <td>0.029590</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                variable      coef\n",
       "4          hemo_difftime -0.111736\n",
       "1       aspirin_timediff -0.061185\n",
       "2         p2y12_timediff -0.033396\n",
       "6             gender_bin -0.027748\n",
       "3  antiplatelet_timediff  0.029590"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coef_tot.sort_values(by='coef').head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4954ab89",
   "metadata": {},
   "source": [
    "There is ?? higher likelihood in delay in hemo treatment,  compared to other features ??"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "3d3c41f8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(coef_tot)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f7484614",
   "metadata": {},
   "source": [
    "### Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "748cda02",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Instantiate models\n",
    "models = {'lr': LogisticRegression(max_iter=1000, solver='saga'),\n",
    "        'dt': DecisionTreeClassifier(),\n",
    "        'rf': RandomForestClassifier(),\n",
    "        'et': ExtraTreesClassifier(),\n",
    "        'svc': SVC(probability=True),\n",
    "        } "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "ee17b45a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_models(model, model_params):\n",
    "    pipe = Pipeline([\n",
    "            ('scale', StandardScaler()),\n",
    "            ('sampling', SMOTE(sampling_strategy = 'minority')),\n",
    "            (model, models[model])\n",
    "            ])\n",
    "    \n",
    "    grid = GridSearchCV(pipe, param_grid=model_params, cv=3, scoring='roc_auc', n_jobs=-1)\n",
    "    grid.fit(X_train, y_train)\n",
    "\n",
    "    # metrics\n",
    "    train_score = grid.score(X_train, y_train)\n",
    "    test_score = grid.score(X_test, y_test)\n",
    "    preds = grid.predict(X_test)\n",
    "    tn, fp, fn, tp = confusion_matrix(y_test, preds).ravel()\n",
    "  \n",
    "    # View confusion matrix\n",
    "    plot_confusion_matrix(grid, X_test, y_test, cmap='Blues', values_format='d');\n",
    "    \n",
    "    # Calculate the sensitivity/ recall\n",
    "    sens = tp / (tp + fn)\n",
    "    \n",
    "    # Calculate the specificity\n",
    "    spec = tn / (tn + fp) \n",
    "    \n",
    "    # print results\n",
    "    print(f'Best params: {grid.best_params_}')\n",
    "    print(f'Training score: {round(train_score, 4)}')\n",
    "    print(f'Test score: {round(test_score, 4)}')\n",
    "    print(f'Recall: {round(sens, 4)}')\n",
    "    print(f'Specificity: {round(spec, 4)}')\n",
    "    return grid"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d4cf27cf",
   "metadata": {},
   "source": [
    "**Logistic Regression**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "8b30631d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best params: {'lr__C': 0.1, 'lr__penalty': 'l1', 'sampling__k_neighbors': 3}\n",
      "Training score: 0.7443\n",
      "Test score: 0.7293\n",
      "Recall: 0.6354\n",
      "Specificity: 0.7291\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# logistic regression\n",
    "grid_params = {\n",
    "    'sampling__k_neighbors': [1,2,3],  \n",
    "    'lr__penalty': ['l1', 'l2',  'elasticnet'],   \n",
    "    'lr__C': [0.1, 1]  \n",
    "} \n",
    "\n",
    "lr = run_models('lr', grid_params)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b5e9ad18",
   "metadata": {},
   "source": [
    "**Decision Tree Classifier**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "2ee47063",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best params: {'dt__max_depth': 7, 'dt__min_samples_leaf': 8, 'dt__min_samples_split': 8, 'sampling__k_neighbors': 5}\n",
      "Training score: 0.7849\n",
      "Test score: 0.7207\n",
      "Recall: 0.5234\n",
      "Specificity: 0.7977\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# decision tree classifier\n",
    "grid_params = {\n",
    "    'sampling__k_neighbors': [4,5,6],  \n",
    "    'dt__max_depth': [7,8,9],\n",
    "    'dt__min_samples_split': [7,8,9],\n",
    "    'dt__min_samples_leaf': [7,8,9],\n",
    "} \n",
    "\n",
    "dt = run_models('dt', grid_params)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "635043b7",
   "metadata": {},
   "source": [
    "**Random forest classifier**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "76cebe3d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best params: {'rf__max_depth': 5, 'rf__min_samples_leaf': 8, 'rf__min_samples_split': 7, 'rf__n_estimators': 8, 'sampling__k_neighbors': 7}\n",
      "Training score: 0.7859\n",
      "Test score: 0.7495\n",
      "Recall: 0.6146\n",
      "Specificity: 0.7603\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# random forest classifier\n",
    "grid_params = {\n",
    "    'sampling__k_neighbors': [7,8,9],\n",
    "    'rf__n_estimators': [7,8,9],\n",
    "    'rf__max_depth': [5, 8, 9],\n",
    "    'rf__min_samples_split': [7,8,9],\n",
    "    'rf__min_samples_leaf': [7,8,9],\n",
    "} \n",
    "\n",
    "rf = run_models('rf', grid_params)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1aa4e493",
   "metadata": {},
   "source": [
    "**Extra trees classifier**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "476a1a73",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best params: {'et__max_depth': 6, 'et__min_samples_leaf': 15, 'et__min_samples_split': 15, 'et__n_estimators': 30, 'sampling__k_neighbors': 20}\n",
      "Training score: 0.7738\n",
      "Test score: 0.7536\n",
      "Recall: 0.6406\n",
      "Specificity: 0.7386\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# extra trees classifier\n",
    "grid_params = {\n",
    "    'sampling__k_neighbors': [15, 20, 25],\n",
    "    'et__n_estimators': [30, 35],\n",
    "    'et__max_depth': [4, 5, 6],\n",
    "    'et__min_samples_split': [10, 15],\n",
    "    'et__min_samples_leaf': [10, 15],\n",
    "} \n",
    "\n",
    "et = run_models('et', grid_params)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a770133a",
   "metadata": {},
   "source": [
    "**Feature importance**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e61f2f74",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "655131cc",
   "metadata": {},
   "source": [
    "## Plot ROC-AUC "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "671f713b",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_all_auc(models):\n",
    "    fig, ax = plt.subplots(figsize=(10,10))\n",
    "    axes = {}\n",
    "    for i, m in enumerate(models.keys()):\n",
    "        axes[f'ax{i}'] = plot_roc_curve(m, X_test, y_test, ax=ax, name=models[m])\n",
    "        \n",
    "    plt.plot([0, 1], [0, 1], color='k', linestyle='--')\n",
    "    plt.title('ROC-AUC curve',fontsize=20)\n",
    "    plt.xlabel('False Positive Rate (FPR)',fontsize=16)\n",
    "    plt.ylabel('True Positive Rate (TPR)',fontsize=16)\n",
    "    plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "b60b92bf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "models_built = {\n",
    "    lr : 'LogisticRegression', \n",
    "    dt : 'DecisionTreeClassifier',\n",
    "    rf : 'RandomForestClassifier',\n",
    "    et : 'ExtraTreesClassifier',\n",
    "} \n",
    "\n",
    "plot_all_auc(models_built)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "70f6b5bb",
   "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"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "170.333px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}