main.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "df1bddad",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:08:45.481528Z",
     "iopub.status.busy": "2025-01-21T13:08:45.481280Z",
     "iopub.status.idle": "2025-01-21T13:09:05.636222Z",
     "shell.execute_reply": "2025-01-21T13:09:05.635115Z"
    },
    "papermill": {
     "duration": 20.169611,
     "end_time": "2025-01-21T13:09:05.637839",
     "exception": false,
     "start_time": "2025-01-21T13:08:45.468228",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing /kaggle/input/pip-install-lifelines/autograd-1.7.0-py3-none-any.whl\r\n",
      "Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from autograd==1.7.0) (1.26.4)\r\n",
      "autograd is already installed with the same version as the provided wheel. Use --force-reinstall to force an installation of the wheel.\r\n",
      "Processing /kaggle/input/pip-install-lifelines/autograd-gamma-0.5.0.tar.gz\r\n",
      "  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\r\n",
      "Requirement already satisfied: autograd>=1.2.0 in /usr/local/lib/python3.10/dist-packages (from autograd-gamma==0.5.0) (1.7.0)\r\n",
      "Requirement already satisfied: scipy>=1.2.0 in /usr/local/lib/python3.10/dist-packages (from autograd-gamma==0.5.0) (1.13.1)\r\n",
      "Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from autograd>=1.2.0->autograd-gamma==0.5.0) (1.26.4)\r\n",
      "Building wheels for collected packages: autograd-gamma\r\n",
      "  Building wheel for autograd-gamma (setup.py) ... \u001b[?25l\u001b[?25hdone\r\n",
      "  Created wheel for autograd-gamma: filename=autograd_gamma-0.5.0-py3-none-any.whl size=4031 sha256=f747b9bdf014658bd71e0cd71c90d711c22062e71130b9b4988fb99a1f6fcc84\r\n",
      "  Stored in directory: /root/.cache/pip/wheels/6b/b5/e0/4c79e15c0b5f2c15ecf613c720bb20daab20a666eb67135155\r\n",
      "Successfully built autograd-gamma\r\n",
      "Installing collected packages: autograd-gamma\r\n",
      "Successfully installed autograd-gamma-0.5.0\r\n",
      "Processing /kaggle/input/pip-install-lifelines/interface_meta-1.3.0-py3-none-any.whl\r\n",
      "Installing collected packages: interface-meta\r\n",
      "Successfully installed interface-meta-1.3.0\r\n",
      "Processing /kaggle/input/pip-install-lifelines/formulaic-1.0.2-py3-none-any.whl\r\n",
      "Requirement already satisfied: interface-meta>=1.2.0 in /usr/local/lib/python3.10/dist-packages (from formulaic==1.0.2) (1.3.0)\r\n",
      "Requirement already satisfied: numpy>=1.16.5 in /usr/local/lib/python3.10/dist-packages (from formulaic==1.0.2) (1.26.4)\r\n",
      "Requirement already satisfied: pandas>=1.0 in /usr/local/lib/python3.10/dist-packages (from formulaic==1.0.2) (2.1.4)\r\n",
      "Requirement already satisfied: scipy>=1.6 in /usr/local/lib/python3.10/dist-packages (from formulaic==1.0.2) (1.13.1)\r\n",
      "Requirement already satisfied: typing-extensions>=4.2.0 in /usr/local/lib/python3.10/dist-packages (from formulaic==1.0.2) (4.12.2)\r\n",
      "Requirement already satisfied: wrapt>=1.0 in /usr/local/lib/python3.10/dist-packages (from formulaic==1.0.2) (1.16.0)\r\n",
      "Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.10/dist-packages (from pandas>=1.0->formulaic==1.0.2) (2.8.2)\r\n",
      "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas>=1.0->formulaic==1.0.2) (2024.2)\r\n",
      "Requirement already satisfied: tzdata>=2022.1 in /usr/local/lib/python3.10/dist-packages (from pandas>=1.0->formulaic==1.0.2) (2024.1)\r\n",
      "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.8.2->pandas>=1.0->formulaic==1.0.2) (1.16.0)\r\n",
      "Installing collected packages: formulaic\r\n",
      "Successfully installed formulaic-1.0.2\r\n",
      "Processing /kaggle/input/pip-install-lifelines/lifelines-0.30.0-py3-none-any.whl\r\n",
      "Requirement already satisfied: numpy>=1.14.0 in /usr/local/lib/python3.10/dist-packages (from lifelines==0.30.0) (1.26.4)\r\n",
      "Requirement already satisfied: scipy>=1.7.0 in /usr/local/lib/python3.10/dist-packages (from lifelines==0.30.0) (1.13.1)\r\n",
      "Requirement already satisfied: pandas>=2.1 in /usr/local/lib/python3.10/dist-packages (from lifelines==0.30.0) (2.1.4)\r\n",
      "Requirement already satisfied: matplotlib>=3.0 in /usr/local/lib/python3.10/dist-packages (from lifelines==0.30.0) (3.7.1)\r\n",
      "Requirement already satisfied: autograd>=1.5 in /usr/local/lib/python3.10/dist-packages (from lifelines==0.30.0) (1.7.0)\r\n",
      "Requirement already satisfied: autograd-gamma>=0.3 in /usr/local/lib/python3.10/dist-packages (from lifelines==0.30.0) (0.5.0)\r\n",
      "Requirement already satisfied: formulaic>=0.2.2 in /usr/local/lib/python3.10/dist-packages (from lifelines==0.30.0) (1.0.2)\r\n",
      "Requirement already satisfied: interface-meta>=1.2.0 in /usr/local/lib/python3.10/dist-packages (from formulaic>=0.2.2->lifelines==0.30.0) (1.3.0)\r\n",
      "Requirement already satisfied: typing-extensions>=4.2.0 in /usr/local/lib/python3.10/dist-packages (from formulaic>=0.2.2->lifelines==0.30.0) (4.12.2)\r\n",
      "Requirement already satisfied: wrapt>=1.0 in /usr/local/lib/python3.10/dist-packages (from formulaic>=0.2.2->lifelines==0.30.0) (1.16.0)\r\n",
      "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.0->lifelines==0.30.0) (1.3.0)\r\n",
      "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.0->lifelines==0.30.0) (0.12.1)\r\n",
      "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.0->lifelines==0.30.0) (4.53.1)\r\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.0->lifelines==0.30.0) (1.4.7)\r\n",
      "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.0->lifelines==0.30.0) (24.1)\r\n",
      "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.0->lifelines==0.30.0) (10.4.0)\r\n",
      "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.0->lifelines==0.30.0) (3.1.4)\r\n",
      "Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib>=3.0->lifelines==0.30.0) (2.8.2)\r\n",
      "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas>=2.1->lifelines==0.30.0) (2024.2)\r\n",
      "Requirement already satisfied: tzdata>=2022.1 in /usr/local/lib/python3.10/dist-packages (from pandas>=2.1->lifelines==0.30.0) (2024.1)\r\n",
      "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.7->matplotlib>=3.0->lifelines==0.30.0) (1.16.0)\r\n",
      "Installing collected packages: lifelines\r\n",
      "Successfully installed lifelines-0.30.0\r\n"
     ]
    }
   ],
   "source": [
    "!pip install /kaggle/input/pip-install-lifelines/autograd-1.7.0-py3-none-any.whl\n",
    "!pip install /kaggle/input/pip-install-lifelines/autograd-gamma-0.5.0.tar.gz\n",
    "!pip install /kaggle/input/pip-install-lifelines/interface_meta-1.3.0-py3-none-any.whl\n",
    "!pip install /kaggle/input/pip-install-lifelines/formulaic-1.0.2-py3-none-any.whl\n",
    "!pip install /kaggle/input/pip-install-lifelines/lifelines-0.30.0-py3-none-any.whl"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e3f63e69",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:05.664650Z",
     "iopub.status.busy": "2025-01-21T13:09:05.664364Z",
     "iopub.status.idle": "2025-01-21T13:09:07.449833Z",
     "shell.execute_reply": "2025-01-21T13:09:07.448915Z"
    },
    "papermill": {
     "duration": 1.800534,
     "end_time": "2025-01-21T13:09:07.451647",
     "exception": false,
     "start_time": "2025-01-21T13:09:05.651113",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "To evaluate the equitable prediction of transplant survival outcomes,\n",
    "we use the concordance index (C-index) between a series of event\n",
    "times and a predicted score across each race group.\n",
    " \n",
    "It represents the global assessment of the model discrimination power:\n",
    "this is the model’s ability to correctly provide a reliable ranking\n",
    "of the survival times based on the individual risk scores.\n",
    " \n",
    "The concordance index is a value between 0 and 1 where:\n",
    " \n",
    "0.5 is the expected result from random predictions,\n",
    "1.0 is perfect concordance (with no censoring, otherwise <1.0),\n",
    "0.0 is perfect anti-concordance (with no censoring, otherwise >0.0)\n",
    "\n",
    "\"\"\"\n",
    "\n",
    "import pandas as pd\n",
    "import pandas.api.types\n",
    "import numpy as np\n",
    "from lifelines.utils import concordance_index\n",
    "\n",
    "class ParticipantVisibleError(Exception):\n",
    "    pass\n",
    "\n",
    "\n",
    "def score(solution: pd.DataFrame, submission: pd.DataFrame, row_id_column_name: str) -> float:\n",
    "    \"\"\"\n",
    "    >>> import pandas as pd\n",
    "    >>> row_id_column_name = \"id\"\n",
    "    >>> y_pred = {'prediction': {0: 1.0, 1: 0.0, 2: 1.0}}\n",
    "    >>> y_pred = pd.DataFrame(y_pred)\n",
    "    >>> y_pred.insert(0, row_id_column_name, range(len(y_pred)))\n",
    "    >>> y_true = { 'efs': {0: 1.0, 1: 0.0, 2: 0.0}, 'efs_time': {0: 25.1234,1: 250.1234,2: 2500.1234}, 'race_group': {0: 'race_group_1', 1: 'race_group_1', 2: 'race_group_1'}}\n",
    "    >>> y_true = pd.DataFrame(y_true)\n",
    "    >>> y_true.insert(0, row_id_column_name, range(len(y_true)))\n",
    "    >>> score(y_true.copy(), y_pred.copy(), row_id_column_name)\n",
    "    0.75\n",
    "    \"\"\"\n",
    "    \n",
    "    del solution[row_id_column_name]\n",
    "    del submission[row_id_column_name]\n",
    "    \n",
    "    event_label = 'efs'\n",
    "    interval_label = 'efs_time'\n",
    "    prediction_label = 'prediction'\n",
    "    for col in submission.columns:\n",
    "        if not pandas.api.types.is_numeric_dtype(submission[col]):\n",
    "            raise ParticipantVisibleError(f'Submission column {col} must be a number')\n",
    "    # Merging solution and submission dfs on ID\n",
    "    merged_df = pd.concat([solution, submission], axis=1)\n",
    "    merged_df.reset_index(inplace=True)\n",
    "    merged_df_race_dict = dict(merged_df.groupby(['race_group']).groups)\n",
    "    metric_list = []\n",
    "    for race in merged_df_race_dict.keys():\n",
    "        # Retrieving values from y_test based on index\n",
    "        indices = sorted(merged_df_race_dict[race])\n",
    "        merged_df_race = merged_df.iloc[indices]\n",
    "        # Calculate the concordance index\n",
    "        c_index_race = concordance_index(\n",
    "                        merged_df_race[interval_label],\n",
    "                        -merged_df_race[prediction_label],\n",
    "                        merged_df_race[event_label])\n",
    "        metric_list.append(c_index_race)\n",
    "    return float(np.mean(metric_list)-np.sqrt(np.var(metric_list)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "4dca53da",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:07.478419Z",
     "iopub.status.busy": "2025-01-21T13:09:07.478058Z",
     "iopub.status.idle": "2025-01-21T13:09:07.481730Z",
     "shell.execute_reply": "2025-01-21T13:09:07.480957Z"
    },
    "papermill": {
     "duration": 0.01792,
     "end_time": "2025-01-21T13:09:07.483085",
     "exception": false,
     "start_time": "2025-01-21T13:09:07.465165",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import numpy as np, pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "pd.set_option('display.max_columns', 500)\n",
    "pd.set_option('display.max_rows', 500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "643d9bd5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:07.510470Z",
     "iopub.status.busy": "2025-01-21T13:09:07.510257Z",
     "iopub.status.idle": "2025-01-21T13:09:07.859533Z",
     "shell.execute_reply": "2025-01-21T13:09:07.858787Z"
    },
    "papermill": {
     "duration": 0.365028,
     "end_time": "2025-01-21T13:09:07.860884",
     "exception": false,
     "start_time": "2025-01-21T13:09:07.495856",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test shape: (3, 58)\n",
      "Train shape: (28800, 60)\n"
     ]
    }
   ],
   "source": [
    "test = pd.read_csv(\"/kaggle/input/equity-post-HCT-survival-predictions/test.csv\")\n",
    "print(\"Test shape:\", test.shape )\n",
    "\n",
    "\n",
    "train = pd.read_csv(\"/kaggle/input/equity-post-HCT-survival-predictions/train.csv\")\n",
    "print(\"Train shape:\",train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "bd01efff",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:07.887210Z",
     "iopub.status.busy": "2025-01-21T13:09:07.886978Z",
     "iopub.status.idle": "2025-01-21T13:09:07.934953Z",
     "shell.execute_reply": "2025-01-21T13:09:07.934111Z"
    },
    "papermill": {
     "duration": 0.062874,
     "end_time": "2025-01-21T13:09:07.936498",
     "exception": false,
     "start_time": "2025-01-21T13:09:07.873624",
     "status": "completed"
    },
    "tags": []
   },
   "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>ID</th>\n",
       "      <th>dri_score</th>\n",
       "      <th>psych_disturb</th>\n",
       "      <th>cyto_score</th>\n",
       "      <th>diabetes</th>\n",
       "      <th>hla_match_c_high</th>\n",
       "      <th>hla_high_res_8</th>\n",
       "      <th>tbi_status</th>\n",
       "      <th>arrhythmia</th>\n",
       "      <th>hla_low_res_6</th>\n",
       "      <th>graft_type</th>\n",
       "      <th>vent_hist</th>\n",
       "      <th>renal_issue</th>\n",
       "      <th>pulm_severe</th>\n",
       "      <th>prim_disease_hct</th>\n",
       "      <th>hla_high_res_6</th>\n",
       "      <th>cmv_status</th>\n",
       "      <th>hla_high_res_10</th>\n",
       "      <th>hla_match_dqb1_high</th>\n",
       "      <th>tce_imm_match</th>\n",
       "      <th>hla_nmdp_6</th>\n",
       "      <th>hla_match_c_low</th>\n",
       "      <th>rituximab</th>\n",
       "      <th>hla_match_drb1_low</th>\n",
       "      <th>hla_match_dqb1_low</th>\n",
       "      <th>prod_type</th>\n",
       "      <th>cyto_score_detail</th>\n",
       "      <th>conditioning_intensity</th>\n",
       "      <th>ethnicity</th>\n",
       "      <th>year_hct</th>\n",
       "      <th>obesity</th>\n",
       "      <th>mrd_hct</th>\n",
       "      <th>in_vivo_tcd</th>\n",
       "      <th>tce_match</th>\n",
       "      <th>hla_match_a_high</th>\n",
       "      <th>hepatic_severe</th>\n",
       "      <th>donor_age</th>\n",
       "      <th>prior_tumor</th>\n",
       "      <th>hla_match_b_low</th>\n",
       "      <th>peptic_ulcer</th>\n",
       "      <th>age_at_hct</th>\n",
       "      <th>hla_match_a_low</th>\n",
       "      <th>gvhd_proph</th>\n",
       "      <th>rheum_issue</th>\n",
       "      <th>sex_match</th>\n",
       "      <th>hla_match_b_high</th>\n",
       "      <th>race_group</th>\n",
       "      <th>comorbidity_score</th>\n",
       "      <th>karnofsky_score</th>\n",
       "      <th>hepatic_mild</th>\n",
       "      <th>tce_div_match</th>\n",
       "      <th>donor_related</th>\n",
       "      <th>melphalan_dose</th>\n",
       "      <th>hla_low_res_8</th>\n",
       "      <th>cardiac</th>\n",
       "      <th>hla_match_drb1_high</th>\n",
       "      <th>pulm_moderate</th>\n",
       "      <th>hla_low_res_10</th>\n",
       "      <th>efs</th>\n",
       "      <th>efs_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>N/A - non-malignant indication</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No TBI</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Bone marrow</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>IEA</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>BM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2016</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>9.942</td>\n",
       "      <td>2.0</td>\n",
       "      <td>FKalone</td>\n",
       "      <td>No</td>\n",
       "      <td>M-F</td>\n",
       "      <td>2.0</td>\n",
       "      <td>More than one race</td>\n",
       "      <td>0.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unrelated</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>42.356</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>No</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>TBI +- Other, &gt;cGy</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Peripheral blood</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>AML</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>P/P</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>PB</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>MAC</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2008</td>\n",
       "      <td>No</td>\n",
       "      <td>Positive</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>72.29</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>43.705</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Other GVHD Prophylaxis</td>\n",
       "      <td>No</td>\n",
       "      <td>F-F</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Asian</td>\n",
       "      <td>3.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive mismatched</td>\n",
       "      <td>Related</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>10.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>N/A - non-malignant indication</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No TBI</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Bone marrow</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>HIS</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>P/P</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>BM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2019</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>33.997</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Cyclophosphamide alone</td>\n",
       "      <td>No</td>\n",
       "      <td>F-M</td>\n",
       "      <td>2.0</td>\n",
       "      <td>More than one race</td>\n",
       "      <td>0.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive mismatched</td>\n",
       "      <td>Related</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19.793</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>High</td>\n",
       "      <td>No</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No TBI</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Bone marrow</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>ALL</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>P/P</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>BM</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>MAC</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2009</td>\n",
       "      <td>No</td>\n",
       "      <td>Positive</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>29.23</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>43.245</td>\n",
       "      <td>2.0</td>\n",
       "      <td>FK+ MMF +- others</td>\n",
       "      <td>No</td>\n",
       "      <td>M-M</td>\n",
       "      <td>2.0</td>\n",
       "      <td>White</td>\n",
       "      <td>0.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Permissive mismatched</td>\n",
       "      <td>Unrelated</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>102.349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>High</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No TBI</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Peripheral blood</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>MPN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>PB</td>\n",
       "      <td>NaN</td>\n",
       "      <td>MAC</td>\n",
       "      <td>Hispanic or Latino</td>\n",
       "      <td>2018</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>56.81</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>29.740</td>\n",
       "      <td>2.0</td>\n",
       "      <td>TDEPLETION +- other</td>\n",
       "      <td>No</td>\n",
       "      <td>M-F</td>\n",
       "      <td>2.0</td>\n",
       "      <td>American Indian or Alaska Native</td>\n",
       "      <td>1.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive mismatched</td>\n",
       "      <td>Related</td>\n",
       "      <td>MEL</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>16.223</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   ID                       dri_score psych_disturb    cyto_score diabetes  \\\n",
       "0   0  N/A - non-malignant indication            No           NaN       No   \n",
       "1   1                    Intermediate            No  Intermediate       No   \n",
       "2   2  N/A - non-malignant indication            No           NaN       No   \n",
       "3   3                            High            No  Intermediate       No   \n",
       "4   4                            High            No           NaN       No   \n",
       "\n",
       "   hla_match_c_high  hla_high_res_8          tbi_status arrhythmia  \\\n",
       "0               NaN             NaN              No TBI         No   \n",
       "1               2.0             8.0  TBI +- Other, >cGy         No   \n",
       "2               2.0             8.0              No TBI         No   \n",
       "3               2.0             8.0              No TBI         No   \n",
       "4               2.0             8.0              No TBI         No   \n",
       "\n",
       "   hla_low_res_6        graft_type vent_hist renal_issue pulm_severe  \\\n",
       "0            6.0       Bone marrow        No          No          No   \n",
       "1            6.0  Peripheral blood        No          No          No   \n",
       "2            6.0       Bone marrow        No          No          No   \n",
       "3            6.0       Bone marrow        No          No          No   \n",
       "4            6.0  Peripheral blood        No          No          No   \n",
       "\n",
       "  prim_disease_hct  hla_high_res_6 cmv_status  hla_high_res_10  \\\n",
       "0              IEA             6.0        +/+              NaN   \n",
       "1              AML             6.0        +/+             10.0   \n",
       "2              HIS             6.0        +/+             10.0   \n",
       "3              ALL             6.0        +/+             10.0   \n",
       "4              MPN             6.0        +/+             10.0   \n",
       "\n",
       "   hla_match_dqb1_high tce_imm_match  hla_nmdp_6  hla_match_c_low rituximab  \\\n",
       "0                  2.0           NaN         6.0              2.0        No   \n",
       "1                  2.0           P/P         6.0              2.0        No   \n",
       "2                  2.0           P/P         6.0              2.0        No   \n",
       "3                  2.0           P/P         6.0              2.0        No   \n",
       "4                  2.0           NaN         5.0              2.0        No   \n",
       "\n",
       "   hla_match_drb1_low  hla_match_dqb1_low prod_type cyto_score_detail  \\\n",
       "0                 2.0                 2.0        BM               NaN   \n",
       "1                 2.0                 2.0        PB      Intermediate   \n",
       "2                 2.0                 2.0        BM               NaN   \n",
       "3                 2.0                 2.0        BM      Intermediate   \n",
       "4                 2.0                 2.0        PB               NaN   \n",
       "\n",
       "  conditioning_intensity               ethnicity  year_hct obesity   mrd_hct  \\\n",
       "0                    NaN  Not Hispanic or Latino      2016      No       NaN   \n",
       "1                    MAC  Not Hispanic or Latino      2008      No  Positive   \n",
       "2                    NaN  Not Hispanic or Latino      2019      No       NaN   \n",
       "3                    MAC  Not Hispanic or Latino      2009      No  Positive   \n",
       "4                    MAC      Hispanic or Latino      2018      No       NaN   \n",
       "\n",
       "  in_vivo_tcd   tce_match  hla_match_a_high hepatic_severe  donor_age  \\\n",
       "0         Yes         NaN               2.0             No        NaN   \n",
       "1          No  Permissive               2.0             No      72.29   \n",
       "2         Yes         NaN               2.0             No        NaN   \n",
       "3          No  Permissive               2.0             No      29.23   \n",
       "4         Yes         NaN               2.0             No      56.81   \n",
       "\n",
       "  prior_tumor  hla_match_b_low peptic_ulcer  age_at_hct  hla_match_a_low  \\\n",
       "0          No              2.0           No       9.942              2.0   \n",
       "1          No              2.0           No      43.705              2.0   \n",
       "2          No              2.0           No      33.997              2.0   \n",
       "3          No              2.0           No      43.245              2.0   \n",
       "4          No              2.0           No      29.740              2.0   \n",
       "\n",
       "               gvhd_proph rheum_issue sex_match  hla_match_b_high  \\\n",
       "0                 FKalone          No       M-F               2.0   \n",
       "1  Other GVHD Prophylaxis          No       F-F               2.0   \n",
       "2  Cyclophosphamide alone          No       F-M               2.0   \n",
       "3       FK+ MMF +- others          No       M-M               2.0   \n",
       "4     TDEPLETION +- other          No       M-F               2.0   \n",
       "\n",
       "                         race_group  comorbidity_score  karnofsky_score  \\\n",
       "0                More than one race                0.0             90.0   \n",
       "1                             Asian                3.0             90.0   \n",
       "2                More than one race                0.0             90.0   \n",
       "3                             White                0.0             90.0   \n",
       "4  American Indian or Alaska Native                1.0             90.0   \n",
       "\n",
       "  hepatic_mild          tce_div_match donor_related      melphalan_dose  \\\n",
       "0           No                    NaN     Unrelated  N/A, Mel not given   \n",
       "1           No  Permissive mismatched       Related  N/A, Mel not given   \n",
       "2           No  Permissive mismatched       Related  N/A, Mel not given   \n",
       "3          Yes  Permissive mismatched     Unrelated  N/A, Mel not given   \n",
       "4           No  Permissive mismatched       Related                 MEL   \n",
       "\n",
       "   hla_low_res_8 cardiac  hla_match_drb1_high pulm_moderate  hla_low_res_10  \\\n",
       "0            8.0      No                  2.0            No            10.0   \n",
       "1            8.0      No                  2.0           Yes            10.0   \n",
       "2            8.0      No                  2.0            No            10.0   \n",
       "3            8.0      No                  2.0            No            10.0   \n",
       "4            8.0      No                  2.0            No            10.0   \n",
       "\n",
       "   efs  efs_time  \n",
       "0  0.0    42.356  \n",
       "1  1.0     4.672  \n",
       "2  0.0    19.793  \n",
       "3  0.0   102.349  \n",
       "4  0.0    16.223  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "98b061e3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:07.963713Z",
     "iopub.status.busy": "2025-01-21T13:09:07.963496Z",
     "iopub.status.idle": "2025-01-21T13:09:07.998759Z",
     "shell.execute_reply": "2025-01-21T13:09:07.997969Z"
    },
    "papermill": {
     "duration": 0.050292,
     "end_time": "2025-01-21T13:09:08.000149",
     "exception": false,
     "start_time": "2025-01-21T13:09:07.949857",
     "status": "completed"
    },
    "tags": []
   },
   "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>ID</th>\n",
       "      <th>dri_score</th>\n",
       "      <th>psych_disturb</th>\n",
       "      <th>cyto_score</th>\n",
       "      <th>diabetes</th>\n",
       "      <th>hla_match_c_high</th>\n",
       "      <th>hla_high_res_8</th>\n",
       "      <th>tbi_status</th>\n",
       "      <th>arrhythmia</th>\n",
       "      <th>hla_low_res_6</th>\n",
       "      <th>graft_type</th>\n",
       "      <th>vent_hist</th>\n",
       "      <th>renal_issue</th>\n",
       "      <th>pulm_severe</th>\n",
       "      <th>prim_disease_hct</th>\n",
       "      <th>hla_high_res_6</th>\n",
       "      <th>cmv_status</th>\n",
       "      <th>hla_high_res_10</th>\n",
       "      <th>hla_match_dqb1_high</th>\n",
       "      <th>tce_imm_match</th>\n",
       "      <th>hla_nmdp_6</th>\n",
       "      <th>hla_match_c_low</th>\n",
       "      <th>rituximab</th>\n",
       "      <th>hla_match_drb1_low</th>\n",
       "      <th>hla_match_dqb1_low</th>\n",
       "      <th>prod_type</th>\n",
       "      <th>cyto_score_detail</th>\n",
       "      <th>conditioning_intensity</th>\n",
       "      <th>ethnicity</th>\n",
       "      <th>year_hct</th>\n",
       "      <th>obesity</th>\n",
       "      <th>mrd_hct</th>\n",
       "      <th>in_vivo_tcd</th>\n",
       "      <th>tce_match</th>\n",
       "      <th>hla_match_a_high</th>\n",
       "      <th>hepatic_severe</th>\n",
       "      <th>donor_age</th>\n",
       "      <th>prior_tumor</th>\n",
       "      <th>hla_match_b_low</th>\n",
       "      <th>peptic_ulcer</th>\n",
       "      <th>age_at_hct</th>\n",
       "      <th>hla_match_a_low</th>\n",
       "      <th>gvhd_proph</th>\n",
       "      <th>rheum_issue</th>\n",
       "      <th>sex_match</th>\n",
       "      <th>hla_match_b_high</th>\n",
       "      <th>race_group</th>\n",
       "      <th>comorbidity_score</th>\n",
       "      <th>karnofsky_score</th>\n",
       "      <th>hepatic_mild</th>\n",
       "      <th>tce_div_match</th>\n",
       "      <th>donor_related</th>\n",
       "      <th>melphalan_dose</th>\n",
       "      <th>hla_low_res_8</th>\n",
       "      <th>cardiac</th>\n",
       "      <th>hla_match_drb1_high</th>\n",
       "      <th>pulm_moderate</th>\n",
       "      <th>hla_low_res_10</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>28800</td>\n",
       "      <td>N/A - non-malignant indication</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No TBI</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Bone marrow</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>IEA</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>BM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2016</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>9.942</td>\n",
       "      <td>2.0</td>\n",
       "      <td>FKalone</td>\n",
       "      <td>No</td>\n",
       "      <td>M-F</td>\n",
       "      <td>2.0</td>\n",
       "      <td>More than one race</td>\n",
       "      <td>0.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unrelated</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>28801</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>No</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>TBI +- Other, &gt;cGy</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Peripheral blood</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>AML</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>P/P</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>PB</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>MAC</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2008</td>\n",
       "      <td>No</td>\n",
       "      <td>Positive</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>72.29</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>43.705</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Other GVHD Prophylaxis</td>\n",
       "      <td>No</td>\n",
       "      <td>F-F</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Asian</td>\n",
       "      <td>3.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive mismatched</td>\n",
       "      <td>Related</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28802</td>\n",
       "      <td>N/A - non-malignant indication</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No TBI</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Bone marrow</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>HIS</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>P/P</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>BM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2019</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>33.997</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Cyclophosphamide alone</td>\n",
       "      <td>No</td>\n",
       "      <td>F-M</td>\n",
       "      <td>2.0</td>\n",
       "      <td>More than one race</td>\n",
       "      <td>0.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive mismatched</td>\n",
       "      <td>Related</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      ID                       dri_score psych_disturb    cyto_score diabetes  \\\n",
       "0  28800  N/A - non-malignant indication            No           NaN       No   \n",
       "1  28801                    Intermediate            No  Intermediate       No   \n",
       "2  28802  N/A - non-malignant indication            No           NaN       No   \n",
       "\n",
       "   hla_match_c_high  hla_high_res_8          tbi_status arrhythmia  \\\n",
       "0               NaN             NaN              No TBI         No   \n",
       "1               2.0             8.0  TBI +- Other, >cGy         No   \n",
       "2               2.0             8.0              No TBI         No   \n",
       "\n",
       "   hla_low_res_6        graft_type vent_hist renal_issue pulm_severe  \\\n",
       "0            6.0       Bone marrow        No          No          No   \n",
       "1            6.0  Peripheral blood        No          No          No   \n",
       "2            6.0       Bone marrow        No          No          No   \n",
       "\n",
       "  prim_disease_hct  hla_high_res_6 cmv_status  hla_high_res_10  \\\n",
       "0              IEA             6.0        +/+              NaN   \n",
       "1              AML             6.0        +/+             10.0   \n",
       "2              HIS             6.0        +/+             10.0   \n",
       "\n",
       "   hla_match_dqb1_high tce_imm_match  hla_nmdp_6  hla_match_c_low rituximab  \\\n",
       "0                  2.0           NaN         6.0              2.0        No   \n",
       "1                  2.0           P/P         6.0              2.0        No   \n",
       "2                  2.0           P/P         6.0              2.0        No   \n",
       "\n",
       "   hla_match_drb1_low  hla_match_dqb1_low prod_type cyto_score_detail  \\\n",
       "0                 2.0                 2.0        BM               NaN   \n",
       "1                 2.0                 2.0        PB      Intermediate   \n",
       "2                 2.0                 2.0        BM               NaN   \n",
       "\n",
       "  conditioning_intensity               ethnicity  year_hct obesity   mrd_hct  \\\n",
       "0                    NaN  Not Hispanic or Latino      2016      No       NaN   \n",
       "1                    MAC  Not Hispanic or Latino      2008      No  Positive   \n",
       "2                    NaN  Not Hispanic or Latino      2019      No       NaN   \n",
       "\n",
       "  in_vivo_tcd   tce_match  hla_match_a_high hepatic_severe  donor_age  \\\n",
       "0         Yes         NaN               2.0             No        NaN   \n",
       "1          No  Permissive               2.0             No      72.29   \n",
       "2         Yes         NaN               2.0             No        NaN   \n",
       "\n",
       "  prior_tumor  hla_match_b_low peptic_ulcer  age_at_hct  hla_match_a_low  \\\n",
       "0          No              2.0           No       9.942              2.0   \n",
       "1          No              2.0           No      43.705              2.0   \n",
       "2          No              2.0           No      33.997              2.0   \n",
       "\n",
       "               gvhd_proph rheum_issue sex_match  hla_match_b_high  \\\n",
       "0                 FKalone          No       M-F               2.0   \n",
       "1  Other GVHD Prophylaxis          No       F-F               2.0   \n",
       "2  Cyclophosphamide alone          No       F-M               2.0   \n",
       "\n",
       "           race_group  comorbidity_score  karnofsky_score hepatic_mild  \\\n",
       "0  More than one race                0.0             90.0           No   \n",
       "1               Asian                3.0             90.0           No   \n",
       "2  More than one race                0.0             90.0           No   \n",
       "\n",
       "           tce_div_match donor_related      melphalan_dose  hla_low_res_8  \\\n",
       "0                    NaN     Unrelated  N/A, Mel not given            8.0   \n",
       "1  Permissive mismatched       Related  N/A, Mel not given            8.0   \n",
       "2  Permissive mismatched       Related  N/A, Mel not given            8.0   \n",
       "\n",
       "  cardiac  hla_match_drb1_high pulm_moderate  hla_low_res_10  \n",
       "0      No                  2.0            No            10.0  \n",
       "1      No                  2.0           Yes            10.0  \n",
       "2      No                  2.0            No            10.0  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a0a5340a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:08.028516Z",
     "iopub.status.busy": "2025-01-21T13:09:08.028310Z",
     "iopub.status.idle": "2025-01-21T13:09:08.118089Z",
     "shell.execute_reply": "2025-01-21T13:09:08.117278Z"
    },
    "papermill": {
     "duration": 0.105405,
     "end_time": "2025-01-21T13:09:08.119486",
     "exception": false,
     "start_time": "2025-01-21T13:09:08.014081",
     "status": "completed"
    },
    "tags": []
   },
   "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>ID</th>\n",
       "      <th>hla_match_c_high</th>\n",
       "      <th>hla_high_res_8</th>\n",
       "      <th>hla_low_res_6</th>\n",
       "      <th>hla_high_res_6</th>\n",
       "      <th>hla_high_res_10</th>\n",
       "      <th>hla_match_dqb1_high</th>\n",
       "      <th>hla_nmdp_6</th>\n",
       "      <th>hla_match_c_low</th>\n",
       "      <th>hla_match_drb1_low</th>\n",
       "      <th>hla_match_dqb1_low</th>\n",
       "      <th>year_hct</th>\n",
       "      <th>hla_match_a_high</th>\n",
       "      <th>donor_age</th>\n",
       "      <th>hla_match_b_low</th>\n",
       "      <th>age_at_hct</th>\n",
       "      <th>hla_match_a_low</th>\n",
       "      <th>hla_match_b_high</th>\n",
       "      <th>comorbidity_score</th>\n",
       "      <th>karnofsky_score</th>\n",
       "      <th>hla_low_res_8</th>\n",
       "      <th>hla_match_drb1_high</th>\n",
       "      <th>hla_low_res_10</th>\n",
       "      <th>efs</th>\n",
       "      <th>efs_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>28800.000000</td>\n",
       "      <td>24180.000000</td>\n",
       "      <td>22971.000000</td>\n",
       "      <td>25530.000000</td>\n",
       "      <td>23516.000000</td>\n",
       "      <td>21637.000000</td>\n",
       "      <td>23601.000000</td>\n",
       "      <td>24603.000000</td>\n",
       "      <td>26000.000000</td>\n",
       "      <td>26157.000000</td>\n",
       "      <td>24606.000000</td>\n",
       "      <td>28800.000000</td>\n",
       "      <td>24499.000000</td>\n",
       "      <td>26992.000000</td>\n",
       "      <td>26235.000000</td>\n",
       "      <td>28800.000000</td>\n",
       "      <td>26410.000000</td>\n",
       "      <td>24712.00000</td>\n",
       "      <td>28323.000000</td>\n",
       "      <td>27930.00000</td>\n",
       "      <td>25147.000000</td>\n",
       "      <td>25448.000000</td>\n",
       "      <td>23736.000000</td>\n",
       "      <td>28800.000000</td>\n",
       "      <td>28800.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>14399.500000</td>\n",
       "      <td>1.764516</td>\n",
       "      <td>6.876801</td>\n",
       "      <td>5.143322</td>\n",
       "      <td>5.109202</td>\n",
       "      <td>8.617230</td>\n",
       "      <td>1.736876</td>\n",
       "      <td>5.160346</td>\n",
       "      <td>1.757808</td>\n",
       "      <td>1.715296</td>\n",
       "      <td>1.773795</td>\n",
       "      <td>2015.179444</td>\n",
       "      <td>1.703743</td>\n",
       "      <td>42.511591</td>\n",
       "      <td>1.719916</td>\n",
       "      <td>38.663162</td>\n",
       "      <td>1.709087</td>\n",
       "      <td>1.69962</td>\n",
       "      <td>1.702327</td>\n",
       "      <td>83.83208</td>\n",
       "      <td>6.903448</td>\n",
       "      <td>1.707128</td>\n",
       "      <td>8.664687</td>\n",
       "      <td>0.539306</td>\n",
       "      <td>23.237678</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>8313.988213</td>\n",
       "      <td>0.431941</td>\n",
       "      <td>1.564313</td>\n",
       "      <td>1.207757</td>\n",
       "      <td>1.214162</td>\n",
       "      <td>1.905125</td>\n",
       "      <td>0.447687</td>\n",
       "      <td>1.203240</td>\n",
       "      <td>0.435453</td>\n",
       "      <td>0.451282</td>\n",
       "      <td>0.427130</td>\n",
       "      <td>3.153914</td>\n",
       "      <td>0.462213</td>\n",
       "      <td>15.251434</td>\n",
       "      <td>0.454449</td>\n",
       "      <td>21.147581</td>\n",
       "      <td>0.458259</td>\n",
       "      <td>0.46518</td>\n",
       "      <td>1.994443</td>\n",
       "      <td>11.02884</td>\n",
       "      <td>1.565017</td>\n",
       "      <td>0.461179</td>\n",
       "      <td>1.882746</td>\n",
       "      <td>0.498461</td>\n",
       "      <td>24.799748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2008.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>18.010000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.044000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>40.00000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.333000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>7199.750000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2013.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>28.447000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>19.539000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>70.00000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.619750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>14399.500000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2016.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>40.063000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>41.006000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.00000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>90.00000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>9.796500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>21599.250000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2018.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>56.131500</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>55.965250</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.00000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>90.00000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>35.100000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>28799.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2020.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>84.800000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>73.726000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.00000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>100.00000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>156.819000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 ID  hla_match_c_high  hla_high_res_8  hla_low_res_6  \\\n",
       "count  28800.000000      24180.000000    22971.000000   25530.000000   \n",
       "mean   14399.500000          1.764516        6.876801       5.143322   \n",
       "std     8313.988213          0.431941        1.564313       1.207757   \n",
       "min        0.000000          0.000000        2.000000       2.000000   \n",
       "25%     7199.750000          2.000000        6.000000       4.000000   \n",
       "50%    14399.500000          2.000000        8.000000       6.000000   \n",
       "75%    21599.250000          2.000000        8.000000       6.000000   \n",
       "max    28799.000000          2.000000        8.000000       6.000000   \n",
       "\n",
       "       hla_high_res_6  hla_high_res_10  hla_match_dqb1_high    hla_nmdp_6  \\\n",
       "count    23516.000000     21637.000000         23601.000000  24603.000000   \n",
       "mean         5.109202         8.617230             1.736876      5.160346   \n",
       "std          1.214162         1.905125             0.447687      1.203240   \n",
       "min          0.000000         3.000000             0.000000      2.000000   \n",
       "25%          4.000000         7.000000             1.000000      4.000000   \n",
       "50%          6.000000        10.000000             2.000000      6.000000   \n",
       "75%          6.000000        10.000000             2.000000      6.000000   \n",
       "max          6.000000        10.000000             2.000000      6.000000   \n",
       "\n",
       "       hla_match_c_low  hla_match_drb1_low  hla_match_dqb1_low      year_hct  \\\n",
       "count     26000.000000        26157.000000        24606.000000  28800.000000   \n",
       "mean          1.757808            1.715296            1.773795   2015.179444   \n",
       "std           0.435453            0.451282            0.427130      3.153914   \n",
       "min           0.000000            1.000000            0.000000   2008.000000   \n",
       "25%           2.000000            1.000000            2.000000   2013.000000   \n",
       "50%           2.000000            2.000000            2.000000   2016.000000   \n",
       "75%           2.000000            2.000000            2.000000   2018.000000   \n",
       "max           2.000000            2.000000            2.000000   2020.000000   \n",
       "\n",
       "       hla_match_a_high     donor_age  hla_match_b_low    age_at_hct  \\\n",
       "count      24499.000000  26992.000000     26235.000000  28800.000000   \n",
       "mean           1.703743     42.511591         1.719916     38.663162   \n",
       "std            0.462213     15.251434         0.454449     21.147581   \n",
       "min            0.000000     18.010000         0.000000      0.044000   \n",
       "25%            1.000000     28.447000         1.000000     19.539000   \n",
       "50%            2.000000     40.063000         2.000000     41.006000   \n",
       "75%            2.000000     56.131500         2.000000     55.965250   \n",
       "max            2.000000     84.800000         2.000000     73.726000   \n",
       "\n",
       "       hla_match_a_low  hla_match_b_high  comorbidity_score  karnofsky_score  \\\n",
       "count     26410.000000       24712.00000       28323.000000      27930.00000   \n",
       "mean          1.709087           1.69962           1.702327         83.83208   \n",
       "std           0.458259           0.46518           1.994443         11.02884   \n",
       "min           0.000000           0.00000           0.000000         40.00000   \n",
       "25%           1.000000           1.00000           0.000000         70.00000   \n",
       "50%           2.000000           2.00000           1.000000         90.00000   \n",
       "75%           2.000000           2.00000           2.000000         90.00000   \n",
       "max           2.000000           2.00000          10.000000        100.00000   \n",
       "\n",
       "       hla_low_res_8  hla_match_drb1_high  hla_low_res_10           efs  \\\n",
       "count   25147.000000         25448.000000    23736.000000  28800.000000   \n",
       "mean        6.903448             1.707128        8.664687      0.539306   \n",
       "std         1.565017             0.461179        1.882746      0.498461   \n",
       "min         2.000000             0.000000        4.000000      0.000000   \n",
       "25%         6.000000             1.000000        7.000000      0.000000   \n",
       "50%         8.000000             2.000000       10.000000      1.000000   \n",
       "75%         8.000000             2.000000       10.000000      1.000000   \n",
       "max         8.000000             2.000000       10.000000      1.000000   \n",
       "\n",
       "           efs_time  \n",
       "count  28800.000000  \n",
       "mean      23.237678  \n",
       "std       24.799748  \n",
       "min        0.333000  \n",
       "25%        5.619750  \n",
       "50%        9.796500  \n",
       "75%       35.100000  \n",
       "max      156.819000  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "f8ff638c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:08.148292Z",
     "iopub.status.busy": "2025-01-21T13:09:08.148069Z",
     "iopub.status.idle": "2025-01-21T13:09:08.208566Z",
     "shell.execute_reply": "2025-01-21T13:09:08.207333Z"
    },
    "papermill": {
     "duration": 0.076006,
     "end_time": "2025-01-21T13:09:08.209879",
     "exception": false,
     "start_time": "2025-01-21T13:09:08.133873",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 28800 entries, 0 to 28799\n",
      "Data columns (total 60 columns):\n",
      " #   Column                  Non-Null Count  Dtype  \n",
      "---  ------                  --------------  -----  \n",
      " 0   ID                      28800 non-null  int64  \n",
      " 1   dri_score               28646 non-null  object \n",
      " 2   psych_disturb           26738 non-null  object \n",
      " 3   cyto_score              20732 non-null  object \n",
      " 4   diabetes                26681 non-null  object \n",
      " 5   hla_match_c_high        24180 non-null  float64\n",
      " 6   hla_high_res_8          22971 non-null  float64\n",
      " 7   tbi_status              28800 non-null  object \n",
      " 8   arrhythmia              26598 non-null  object \n",
      " 9   hla_low_res_6           25530 non-null  float64\n",
      " 10  graft_type              28800 non-null  object \n",
      " 11  vent_hist               28541 non-null  object \n",
      " 12  renal_issue             26885 non-null  object \n",
      " 13  pulm_severe             26665 non-null  object \n",
      " 14  prim_disease_hct        28800 non-null  object \n",
      " 15  hla_high_res_6          23516 non-null  float64\n",
      " 16  cmv_status              28166 non-null  object \n",
      " 17  hla_high_res_10         21637 non-null  float64\n",
      " 18  hla_match_dqb1_high     23601 non-null  float64\n",
      " 19  tce_imm_match           17667 non-null  object \n",
      " 20  hla_nmdp_6              24603 non-null  float64\n",
      " 21  hla_match_c_low         26000 non-null  float64\n",
      " 22  rituximab               26652 non-null  object \n",
      " 23  hla_match_drb1_low      26157 non-null  float64\n",
      " 24  hla_match_dqb1_low      24606 non-null  float64\n",
      " 25  prod_type               28800 non-null  object \n",
      " 26  cyto_score_detail       16877 non-null  object \n",
      " 27  conditioning_intensity  24011 non-null  object \n",
      " 28  ethnicity               28213 non-null  object \n",
      " 29  year_hct                28800 non-null  int64  \n",
      " 30  obesity                 27040 non-null  object \n",
      " 31  mrd_hct                 12203 non-null  object \n",
      " 32  in_vivo_tcd             28575 non-null  object \n",
      " 33  tce_match               9804 non-null   object \n",
      " 34  hla_match_a_high        24499 non-null  float64\n",
      " 35  hepatic_severe          26929 non-null  object \n",
      " 36  donor_age               26992 non-null  float64\n",
      " 37  prior_tumor             27122 non-null  object \n",
      " 38  hla_match_b_low         26235 non-null  float64\n",
      " 39  peptic_ulcer            26381 non-null  object \n",
      " 40  age_at_hct              28800 non-null  float64\n",
      " 41  hla_match_a_low         26410 non-null  float64\n",
      " 42  gvhd_proph              28575 non-null  object \n",
      " 43  rheum_issue             26617 non-null  object \n",
      " 44  sex_match               28539 non-null  object \n",
      " 45  hla_match_b_high        24712 non-null  float64\n",
      " 46  race_group              28800 non-null  object \n",
      " 47  comorbidity_score       28323 non-null  float64\n",
      " 48  karnofsky_score         27930 non-null  float64\n",
      " 49  hepatic_mild            26883 non-null  object \n",
      " 50  tce_div_match           17404 non-null  object \n",
      " 51  donor_related           28642 non-null  object \n",
      " 52  melphalan_dose          27395 non-null  object \n",
      " 53  hla_low_res_8           25147 non-null  float64\n",
      " 54  cardiac                 26258 non-null  object \n",
      " 55  hla_match_drb1_high     25448 non-null  float64\n",
      " 56  pulm_moderate           26753 non-null  object \n",
      " 57  hla_low_res_10          23736 non-null  float64\n",
      " 58  efs                     28800 non-null  float64\n",
      " 59  efs_time                28800 non-null  float64\n",
      "dtypes: float64(23), int64(2), object(35)\n",
      "memory usage: 13.2+ MB\n"
     ]
    }
   ],
   "source": [
    "train.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d003ffd7",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:08.238297Z",
     "iopub.status.busy": "2025-01-21T13:09:08.238093Z",
     "iopub.status.idle": "2025-01-21T13:09:08.249660Z",
     "shell.execute_reply": "2025-01-21T13:09:08.248828Z"
    },
    "papermill": {
     "duration": 0.027546,
     "end_time": "2025-01-21T13:09:08.251128",
     "exception": false,
     "start_time": "2025-01-21T13:09:08.223582",
     "status": "completed"
    },
    "tags": []
   },
   "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>efs</th>\n",
       "      <th>efs_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>42.356</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>19.793</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>102.349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>16.223</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28795</th>\n",
       "      <td>0.0</td>\n",
       "      <td>18.633</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28796</th>\n",
       "      <td>1.0</td>\n",
       "      <td>4.892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28797</th>\n",
       "      <td>0.0</td>\n",
       "      <td>23.157</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28798</th>\n",
       "      <td>0.0</td>\n",
       "      <td>52.351</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28799</th>\n",
       "      <td>0.0</td>\n",
       "      <td>25.158</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>28800 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       efs  efs_time\n",
       "0      0.0    42.356\n",
       "1      1.0     4.672\n",
       "2      0.0    19.793\n",
       "3      0.0   102.349\n",
       "4      0.0    16.223\n",
       "...    ...       ...\n",
       "28795  0.0    18.633\n",
       "28796  1.0     4.892\n",
       "28797  0.0    23.157\n",
       "28798  0.0    52.351\n",
       "28799  0.0    25.158\n",
       "\n",
       "[28800 rows x 2 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train[['efs', 'efs_time']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "a2b7c852",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:08.284558Z",
     "iopub.status.busy": "2025-01-21T13:09:08.284308Z",
     "iopub.status.idle": "2025-01-21T13:09:08.291958Z",
     "shell.execute_reply": "2025-01-21T13:09:08.291041Z"
    },
    "papermill": {
     "duration": 0.025458,
     "end_time": "2025-01-21T13:09:08.293489",
     "exception": false,
     "start_time": "2025-01-21T13:09:08.268031",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "156.819"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "max(train['efs_time'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "fe8c0618",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:08.322678Z",
     "iopub.status.busy": "2025-01-21T13:09:08.322464Z",
     "iopub.status.idle": "2025-01-21T13:09:08.329217Z",
     "shell.execute_reply": "2025-01-21T13:09:08.328580Z"
    },
    "papermill": {
     "duration": 0.022135,
     "end_time": "2025-01-21T13:09:08.330434",
     "exception": false,
     "start_time": "2025-01-21T13:09:08.308299",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['N/A - non-malignant indication', 'Intermediate', 'High', 'Low',\n",
       "       'N/A - disease not classifiable', 'N/A - pediatric',\n",
       "       'TBD cytogenetics',\n",
       "       'Intermediate - TED AML case <missing cytogenetics', nan,\n",
       "       'High - TED AML case <missing cytogenetics', 'Very high',\n",
       "       'Missing disease status'], dtype=object)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train['dri_score'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f91fff77",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:08.358694Z",
     "iopub.status.busy": "2025-01-21T13:09:08.358465Z",
     "iopub.status.idle": "2025-01-21T13:09:08.986201Z",
     "shell.execute_reply": "2025-01-21T13:09:08.985268Z"
    },
    "papermill": {
     "duration": 0.643417,
     "end_time": "2025-01-21T13:09:08.987756",
     "exception": false,
     "start_time": "2025-01-21T13:09:08.344339",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(train.loc[train.efs==1,\"efs_time\"],bins=100,label=\"efs=1, event occoured\")\n",
    "plt.hist(train.loc[train.efs==0,\"efs_time\"],bins=100,label=\"efs=0, (censored)\")\n",
    "plt.xlabel(\"Time of Observation, efs_time\")\n",
    "plt.ylabel(\"Count\")\n",
    "plt.title(\"Histogram\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "07405eab",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:09.020473Z",
     "iopub.status.busy": "2025-01-21T13:09:09.020253Z",
     "iopub.status.idle": "2025-01-21T13:09:09.297520Z",
     "shell.execute_reply": "2025-01-21T13:09:09.296652Z"
    },
    "papermill": {
     "duration": 0.294334,
     "end_time": "2025-01-21T13:09:09.298806",
     "exception": false,
     "start_time": "2025-01-21T13:09:09.004472",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(train['efs_time'], train['efs'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "4f9a8cee",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:09.329494Z",
     "iopub.status.busy": "2025-01-21T13:09:09.329275Z",
     "iopub.status.idle": "2025-01-21T13:09:09.332801Z",
     "shell.execute_reply": "2025-01-21T13:09:09.332193Z"
    },
    "papermill": {
     "duration": 0.020435,
     "end_time": "2025-01-21T13:09:09.334135",
     "exception": false,
     "start_time": "2025-01-21T13:09:09.313700",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "from lifelines import KaplanMeierFitter\n",
    "\n",
    "def transform_survival_probability(df, time_col='efs_time', event_col='efs'):\n",
    "    \"\"\"\n",
    "    Transform using survival probability estimates\n",
    "    \"\"\"\n",
    "    kmf = KaplanMeierFitter()\n",
    "    kmf.fit(df[time_col], df[event_col])\n",
    "    y = kmf.survival_function_at_times(df[time_col]).values\n",
    "    return y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "8d332bfd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:09.364117Z",
     "iopub.status.busy": "2025-01-21T13:09:09.363866Z",
     "iopub.status.idle": "2025-01-21T13:09:09.413666Z",
     "shell.execute_reply": "2025-01-21T13:09:09.412976Z"
    },
    "papermill": {
     "duration": 0.066054,
     "end_time": "2025-01-21T13:09:09.415104",
     "exception": false,
     "start_time": "2025-01-21T13:09:09.349050",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "train['y'] = transform_survival_probability(train, 'efs_time', 'efs')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "345b3af4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:09.446630Z",
     "iopub.status.busy": "2025-01-21T13:09:09.446407Z",
     "iopub.status.idle": "2025-01-21T13:09:09.485855Z",
     "shell.execute_reply": "2025-01-21T13:09:09.484951Z"
    },
    "papermill": {
     "duration": 0.056693,
     "end_time": "2025-01-21T13:09:09.487161",
     "exception": false,
     "start_time": "2025-01-21T13:09:09.430468",
     "status": "completed"
    },
    "tags": []
   },
   "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>ID</th>\n",
       "      <th>dri_score</th>\n",
       "      <th>psych_disturb</th>\n",
       "      <th>cyto_score</th>\n",
       "      <th>diabetes</th>\n",
       "      <th>hla_match_c_high</th>\n",
       "      <th>hla_high_res_8</th>\n",
       "      <th>tbi_status</th>\n",
       "      <th>arrhythmia</th>\n",
       "      <th>hla_low_res_6</th>\n",
       "      <th>graft_type</th>\n",
       "      <th>vent_hist</th>\n",
       "      <th>renal_issue</th>\n",
       "      <th>pulm_severe</th>\n",
       "      <th>prim_disease_hct</th>\n",
       "      <th>hla_high_res_6</th>\n",
       "      <th>cmv_status</th>\n",
       "      <th>hla_high_res_10</th>\n",
       "      <th>hla_match_dqb1_high</th>\n",
       "      <th>tce_imm_match</th>\n",
       "      <th>hla_nmdp_6</th>\n",
       "      <th>hla_match_c_low</th>\n",
       "      <th>rituximab</th>\n",
       "      <th>hla_match_drb1_low</th>\n",
       "      <th>hla_match_dqb1_low</th>\n",
       "      <th>prod_type</th>\n",
       "      <th>cyto_score_detail</th>\n",
       "      <th>conditioning_intensity</th>\n",
       "      <th>ethnicity</th>\n",
       "      <th>year_hct</th>\n",
       "      <th>obesity</th>\n",
       "      <th>mrd_hct</th>\n",
       "      <th>in_vivo_tcd</th>\n",
       "      <th>tce_match</th>\n",
       "      <th>hla_match_a_high</th>\n",
       "      <th>hepatic_severe</th>\n",
       "      <th>donor_age</th>\n",
       "      <th>prior_tumor</th>\n",
       "      <th>hla_match_b_low</th>\n",
       "      <th>peptic_ulcer</th>\n",
       "      <th>age_at_hct</th>\n",
       "      <th>hla_match_a_low</th>\n",
       "      <th>gvhd_proph</th>\n",
       "      <th>rheum_issue</th>\n",
       "      <th>sex_match</th>\n",
       "      <th>hla_match_b_high</th>\n",
       "      <th>race_group</th>\n",
       "      <th>comorbidity_score</th>\n",
       "      <th>karnofsky_score</th>\n",
       "      <th>hepatic_mild</th>\n",
       "      <th>tce_div_match</th>\n",
       "      <th>donor_related</th>\n",
       "      <th>melphalan_dose</th>\n",
       "      <th>hla_low_res_8</th>\n",
       "      <th>cardiac</th>\n",
       "      <th>hla_match_drb1_high</th>\n",
       "      <th>pulm_moderate</th>\n",
       "      <th>hla_low_res_10</th>\n",
       "      <th>efs</th>\n",
       "      <th>efs_time</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>N/A - non-malignant indication</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No TBI</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Bone marrow</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>IEA</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>BM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2016</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>9.942</td>\n",
       "      <td>2.0</td>\n",
       "      <td>FKalone</td>\n",
       "      <td>No</td>\n",
       "      <td>M-F</td>\n",
       "      <td>2.0</td>\n",
       "      <td>More than one race</td>\n",
       "      <td>0.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Unrelated</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>42.356</td>\n",
       "      <td>0.458687</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>No</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>TBI +- Other, &gt;cGy</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Peripheral blood</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>AML</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>P/P</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>PB</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>MAC</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2008</td>\n",
       "      <td>No</td>\n",
       "      <td>Positive</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>72.29</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>43.705</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Other GVHD Prophylaxis</td>\n",
       "      <td>No</td>\n",
       "      <td>F-F</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Asian</td>\n",
       "      <td>3.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive mismatched</td>\n",
       "      <td>Related</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>10.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.672</td>\n",
       "      <td>0.847759</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>N/A - non-malignant indication</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No TBI</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Bone marrow</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>HIS</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>P/P</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>BM</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2019</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>33.997</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Cyclophosphamide alone</td>\n",
       "      <td>No</td>\n",
       "      <td>F-M</td>\n",
       "      <td>2.0</td>\n",
       "      <td>More than one race</td>\n",
       "      <td>0.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive mismatched</td>\n",
       "      <td>Related</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>19.793</td>\n",
       "      <td>0.462424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>High</td>\n",
       "      <td>No</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No TBI</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Bone marrow</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>ALL</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>P/P</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>BM</td>\n",
       "      <td>Intermediate</td>\n",
       "      <td>MAC</td>\n",
       "      <td>Not Hispanic or Latino</td>\n",
       "      <td>2009</td>\n",
       "      <td>No</td>\n",
       "      <td>Positive</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>29.23</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>43.245</td>\n",
       "      <td>2.0</td>\n",
       "      <td>FK+ MMF +- others</td>\n",
       "      <td>No</td>\n",
       "      <td>M-M</td>\n",
       "      <td>2.0</td>\n",
       "      <td>White</td>\n",
       "      <td>0.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>Yes</td>\n",
       "      <td>Permissive mismatched</td>\n",
       "      <td>Unrelated</td>\n",
       "      <td>N/A, Mel not given</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>102.349</td>\n",
       "      <td>0.456661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>High</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No TBI</td>\n",
       "      <td>No</td>\n",
       "      <td>6.0</td>\n",
       "      <td>Peripheral blood</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>No</td>\n",
       "      <td>MPN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>+/+</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>PB</td>\n",
       "      <td>NaN</td>\n",
       "      <td>MAC</td>\n",
       "      <td>Hispanic or Latino</td>\n",
       "      <td>2018</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Yes</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>56.81</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>29.740</td>\n",
       "      <td>2.0</td>\n",
       "      <td>TDEPLETION +- other</td>\n",
       "      <td>No</td>\n",
       "      <td>M-F</td>\n",
       "      <td>2.0</td>\n",
       "      <td>American Indian or Alaska Native</td>\n",
       "      <td>1.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>No</td>\n",
       "      <td>Permissive mismatched</td>\n",
       "      <td>Related</td>\n",
       "      <td>MEL</td>\n",
       "      <td>8.0</td>\n",
       "      <td>No</td>\n",
       "      <td>2.0</td>\n",
       "      <td>No</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>16.223</td>\n",
       "      <td>0.464674</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   ID                       dri_score psych_disturb    cyto_score diabetes  \\\n",
       "0   0  N/A - non-malignant indication            No           NaN       No   \n",
       "1   1                    Intermediate            No  Intermediate       No   \n",
       "2   2  N/A - non-malignant indication            No           NaN       No   \n",
       "3   3                            High            No  Intermediate       No   \n",
       "4   4                            High            No           NaN       No   \n",
       "\n",
       "   hla_match_c_high  hla_high_res_8          tbi_status arrhythmia  \\\n",
       "0               NaN             NaN              No TBI         No   \n",
       "1               2.0             8.0  TBI +- Other, >cGy         No   \n",
       "2               2.0             8.0              No TBI         No   \n",
       "3               2.0             8.0              No TBI         No   \n",
       "4               2.0             8.0              No TBI         No   \n",
       "\n",
       "   hla_low_res_6        graft_type vent_hist renal_issue pulm_severe  \\\n",
       "0            6.0       Bone marrow        No          No          No   \n",
       "1            6.0  Peripheral blood        No          No          No   \n",
       "2            6.0       Bone marrow        No          No          No   \n",
       "3            6.0       Bone marrow        No          No          No   \n",
       "4            6.0  Peripheral blood        No          No          No   \n",
       "\n",
       "  prim_disease_hct  hla_high_res_6 cmv_status  hla_high_res_10  \\\n",
       "0              IEA             6.0        +/+              NaN   \n",
       "1              AML             6.0        +/+             10.0   \n",
       "2              HIS             6.0        +/+             10.0   \n",
       "3              ALL             6.0        +/+             10.0   \n",
       "4              MPN             6.0        +/+             10.0   \n",
       "\n",
       "   hla_match_dqb1_high tce_imm_match  hla_nmdp_6  hla_match_c_low rituximab  \\\n",
       "0                  2.0           NaN         6.0              2.0        No   \n",
       "1                  2.0           P/P         6.0              2.0        No   \n",
       "2                  2.0           P/P         6.0              2.0        No   \n",
       "3                  2.0           P/P         6.0              2.0        No   \n",
       "4                  2.0           NaN         5.0              2.0        No   \n",
       "\n",
       "   hla_match_drb1_low  hla_match_dqb1_low prod_type cyto_score_detail  \\\n",
       "0                 2.0                 2.0        BM               NaN   \n",
       "1                 2.0                 2.0        PB      Intermediate   \n",
       "2                 2.0                 2.0        BM               NaN   \n",
       "3                 2.0                 2.0        BM      Intermediate   \n",
       "4                 2.0                 2.0        PB               NaN   \n",
       "\n",
       "  conditioning_intensity               ethnicity  year_hct obesity   mrd_hct  \\\n",
       "0                    NaN  Not Hispanic or Latino      2016      No       NaN   \n",
       "1                    MAC  Not Hispanic or Latino      2008      No  Positive   \n",
       "2                    NaN  Not Hispanic or Latino      2019      No       NaN   \n",
       "3                    MAC  Not Hispanic or Latino      2009      No  Positive   \n",
       "4                    MAC      Hispanic or Latino      2018      No       NaN   \n",
       "\n",
       "  in_vivo_tcd   tce_match  hla_match_a_high hepatic_severe  donor_age  \\\n",
       "0         Yes         NaN               2.0             No        NaN   \n",
       "1          No  Permissive               2.0             No      72.29   \n",
       "2         Yes         NaN               2.0             No        NaN   \n",
       "3          No  Permissive               2.0             No      29.23   \n",
       "4         Yes         NaN               2.0             No      56.81   \n",
       "\n",
       "  prior_tumor  hla_match_b_low peptic_ulcer  age_at_hct  hla_match_a_low  \\\n",
       "0          No              2.0           No       9.942              2.0   \n",
       "1          No              2.0           No      43.705              2.0   \n",
       "2          No              2.0           No      33.997              2.0   \n",
       "3          No              2.0           No      43.245              2.0   \n",
       "4          No              2.0           No      29.740              2.0   \n",
       "\n",
       "               gvhd_proph rheum_issue sex_match  hla_match_b_high  \\\n",
       "0                 FKalone          No       M-F               2.0   \n",
       "1  Other GVHD Prophylaxis          No       F-F               2.0   \n",
       "2  Cyclophosphamide alone          No       F-M               2.0   \n",
       "3       FK+ MMF +- others          No       M-M               2.0   \n",
       "4     TDEPLETION +- other          No       M-F               2.0   \n",
       "\n",
       "                         race_group  comorbidity_score  karnofsky_score  \\\n",
       "0                More than one race                0.0             90.0   \n",
       "1                             Asian                3.0             90.0   \n",
       "2                More than one race                0.0             90.0   \n",
       "3                             White                0.0             90.0   \n",
       "4  American Indian or Alaska Native                1.0             90.0   \n",
       "\n",
       "  hepatic_mild          tce_div_match donor_related      melphalan_dose  \\\n",
       "0           No                    NaN     Unrelated  N/A, Mel not given   \n",
       "1           No  Permissive mismatched       Related  N/A, Mel not given   \n",
       "2           No  Permissive mismatched       Related  N/A, Mel not given   \n",
       "3          Yes  Permissive mismatched     Unrelated  N/A, Mel not given   \n",
       "4           No  Permissive mismatched       Related                 MEL   \n",
       "\n",
       "   hla_low_res_8 cardiac  hla_match_drb1_high pulm_moderate  hla_low_res_10  \\\n",
       "0            8.0      No                  2.0            No            10.0   \n",
       "1            8.0      No                  2.0           Yes            10.0   \n",
       "2            8.0      No                  2.0            No            10.0   \n",
       "3            8.0      No                  2.0            No            10.0   \n",
       "4            8.0      No                  2.0            No            10.0   \n",
       "\n",
       "   efs  efs_time         y  \n",
       "0  0.0    42.356  0.458687  \n",
       "1  1.0     4.672  0.847759  \n",
       "2  0.0    19.793  0.462424  \n",
       "3  0.0   102.349  0.456661  \n",
       "4  0.0    16.223  0.464674  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "6fd44af0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:09.519408Z",
     "iopub.status.busy": "2025-01-21T13:09:09.519203Z",
     "iopub.status.idle": "2025-01-21T13:09:09.889832Z",
     "shell.execute_reply": "2025-01-21T13:09:09.888875Z"
    },
    "papermill": {
     "duration": 0.389279,
     "end_time": "2025-01-21T13:09:09.891418",
     "exception": false,
     "start_time": "2025-01-21T13:09:09.502139",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(train['y'],bins=100)\n",
    "plt.xlabel(\"Y value\")\n",
    "plt.ylabel(\"Count\")\n",
    "plt.title(\"Histogram\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "575ee546",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:09.924613Z",
     "iopub.status.busy": "2025-01-21T13:09:09.924346Z",
     "iopub.status.idle": "2025-01-21T13:09:09.928944Z",
     "shell.execute_reply": "2025-01-21T13:09:09.928241Z"
    },
    "papermill": {
     "duration": 0.022174,
     "end_time": "2025-01-21T13:09:09.930228",
     "exception": false,
     "start_time": "2025-01-21T13:09:09.908054",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 57 features\n"
     ]
    }
   ],
   "source": [
    "to_rmv =['ID', 'efs', 'efs_time', 'y']\n",
    "features = [col for col in train.columns if not col in to_rmv]\n",
    "target = 'y'\n",
    "print(f'There are {len(features)} features')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "210c7b15",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:09.962504Z",
     "iopub.status.busy": "2025-01-21T13:09:09.962290Z",
     "iopub.status.idle": "2025-01-21T13:09:09.966815Z",
     "shell.execute_reply": "2025-01-21T13:09:09.966124Z"
    },
    "papermill": {
     "duration": 0.021768,
     "end_time": "2025-01-21T13:09:09.967952",
     "exception": false,
     "start_time": "2025-01-21T13:09:09.946184",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 35 catagorial features\n"
     ]
    }
   ],
   "source": [
    "cat = [c for c in features if train[c].dtype == 'object']\n",
    "print(f'There are {len(cat)} catagorial features')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "9735b9a2",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.000303Z",
     "iopub.status.busy": "2025-01-21T13:09:10.000075Z",
     "iopub.status.idle": "2025-01-21T13:09:10.011769Z",
     "shell.execute_reply": "2025-01-21T13:09:10.011024Z"
    },
    "papermill": {
     "duration": 0.030081,
     "end_time": "2025-01-21T13:09:10.013150",
     "exception": false,
     "start_time": "2025-01-21T13:09:09.983069",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['dri_score', 'psych_disturb', 'cyto_score', 'diabetes', 'tbi_status',\n",
       "       'arrhythmia', 'graft_type', 'vent_hist', 'renal_issue', 'pulm_severe',\n",
       "       'prim_disease_hct', 'cmv_status', 'tce_imm_match', 'rituximab',\n",
       "       'prod_type', 'cyto_score_detail', 'conditioning_intensity', 'ethnicity',\n",
       "       'obesity', 'mrd_hct', 'in_vivo_tcd', 'tce_match', 'hepatic_severe',\n",
       "       'prior_tumor', 'peptic_ulcer', 'gvhd_proph', 'rheum_issue', 'sex_match',\n",
       "       'race_group', 'hepatic_mild', 'tce_div_match', 'donor_related',\n",
       "       'melphalan_dose', 'cardiac', 'pulm_moderate'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train[cat].columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "20f56c7e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.048387Z",
     "iopub.status.busy": "2025-01-21T13:09:10.048161Z",
     "iopub.status.idle": "2025-01-21T13:09:10.052464Z",
     "shell.execute_reply": "2025-01-21T13:09:10.051681Z"
    },
    "papermill": {
     "duration": 0.021638,
     "end_time": "2025-01-21T13:09:10.053642",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.032004",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 22 numerical features\n"
     ]
    }
   ],
   "source": [
    "num = [c for c in features if train[c].dtype != 'object']\n",
    "print(f'There are {len(num)} numerical features')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f328ff7b",
   "metadata": {
    "papermill": {
     "duration": 0.015756,
     "end_time": "2025-01-21T13:09:10.085234",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.069478",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# 1st try: apply label encoder for all cat features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "8d6d71a6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.117386Z",
     "iopub.status.busy": "2025-01-21T13:09:10.117180Z",
     "iopub.status.idle": "2025-01-21T13:09:10.119986Z",
     "shell.execute_reply": "2025-01-21T13:09:10.119324Z"
    },
    "jupyter": {
     "source_hidden": true
    },
    "papermill": {
     "duration": 0.020341,
     "end_time": "2025-01-21T13:09:10.121175",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.100834",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "# encoder = LabelEncoder()\n",
    "\n",
    "# # Apply LabelEncoder only on the training set and then use it on the test set\n",
    "# for column in categorical_columns:\n",
    "#     # Fit the encoder on the training data\n",
    "#     encoder.fit(train[column].astype(str))\n",
    "    \n",
    "#     # Transform both the training and test data using the same encoder\n",
    "#     train[column] = encoder.transform(train[column].astype(str))\n",
    "#     test[column] = encoder.transform(test[column].astype(str))  # Apply transformation on test set as well\n",
    "\n",
    "#     # Ensure the encoded values are integers\n",
    "#     train[column] = train[column].astype('int32')\n",
    "#     test[column] = test[column].astype('int32')\n",
    "\n",
    "#     # Optionally, convert to 'category' dtype if needed for memory efficiency\n",
    "#     train[column] = train[column].astype('category')\n",
    "#     test[column] = test[column].astype('category')\n",
    "\n",
    "# test = test.drop(columns='y')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "210a643b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.153169Z",
     "iopub.status.busy": "2025-01-21T13:09:10.152890Z",
     "iopub.status.idle": "2025-01-21T13:09:10.156019Z",
     "shell.execute_reply": "2025-01-21T13:09:10.155148Z"
    },
    "jupyter": {
     "source_hidden": true
    },
    "papermill": {
     "duration": 0.020307,
     "end_time": "2025-01-21T13:09:10.157301",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.136994",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "86c73183",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.190891Z",
     "iopub.status.busy": "2025-01-21T13:09:10.190658Z",
     "iopub.status.idle": "2025-01-21T13:09:10.193614Z",
     "shell.execute_reply": "2025-01-21T13:09:10.192966Z"
    },
    "jupyter": {
     "source_hidden": true
    },
    "papermill": {
     "duration": 0.021827,
     "end_time": "2025-01-21T13:09:10.195159",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.173332",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# train[features].isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "faeb3f3c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.230320Z",
     "iopub.status.busy": "2025-01-21T13:09:10.230104Z",
     "iopub.status.idle": "2025-01-21T13:09:10.233091Z",
     "shell.execute_reply": "2025-01-21T13:09:10.232385Z"
    },
    "jupyter": {
     "source_hidden": true
    },
    "papermill": {
     "duration": 0.021942,
     "end_time": "2025-01-21T13:09:10.234354",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.212412",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# import warnings\n",
    "# warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "# def convert_64_to_32(df, num_features):\n",
    "#     for c in num_features:\n",
    "#         if df[c].dtype == 'float64':\n",
    "#             df[c] = df[c].astype('float32')\n",
    "#         else:\n",
    "#             df[c] = df[c].astype('int32')\n",
    "#     return df\n",
    "\n",
    "# train = convert_64_to_32(train, num)\n",
    "# test = convert_64_to_32(test, num)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "20e9320a",
   "metadata": {
    "papermill": {
     "duration": 0.016525,
     "end_time": "2025-01-21T13:09:10.266709",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.250184",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Feature eng"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "6c91eba5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.299128Z",
     "iopub.status.busy": "2025-01-21T13:09:10.298924Z",
     "iopub.status.idle": "2025-01-21T13:09:10.396278Z",
     "shell.execute_reply": "2025-01-21T13:09:10.395382Z"
    },
    "papermill": {
     "duration": 0.1152,
     "end_time": "2025-01-21T13:09:10.397581",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.282381",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dri_score\n",
      "Intermediate                                         10436\n",
      "N/A - pediatric                                       4779\n",
      "High                                                  4701\n",
      "N/A - non-malignant indication                        2427\n",
      "TBD cytogenetics                                      2003\n",
      "Low                                                   1926\n",
      "High - TED AML case <missing cytogenetics             1414\n",
      "Intermediate - TED AML case <missing cytogenetics      481\n",
      "N/A - disease not classifiable                         272\n",
      "Very high                                              198\n",
      "Missing disease status                                   9\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "psych_disturb\n",
      "No          23005\n",
      "Yes          3587\n",
      "Not done      146\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "cyto_score\n",
      "Poor            8802\n",
      "Intermediate    6376\n",
      "Favorable       3011\n",
      "TBD             1341\n",
      "Normal           643\n",
      "Other            504\n",
      "Not tested        55\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "diabetes\n",
      "No          22201\n",
      "Yes          4339\n",
      "Not done      141\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "tbi_status\n",
      "No TBI                              18861\n",
      "TBI + Cy +- Other                    6104\n",
      "TBI +- Other, <=cGy                  1727\n",
      "TBI +- Other, >cGy                   1700\n",
      "TBI +- Other, -cGy, single            134\n",
      "TBI +- Other, -cGy, fractionated      119\n",
      "TBI +- Other, -cGy, unknown dose       79\n",
      "TBI +- Other, unknown dose             76\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "arrhythmia\n",
      "No          25203\n",
      "Yes          1277\n",
      "Not done      118\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "graft_type\n",
      "Peripheral blood    20546\n",
      "Bone marrow          8254\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "vent_hist\n",
      "No     27721\n",
      "Yes      820\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "renal_issue\n",
      "No          26548\n",
      "Yes           200\n",
      "Not done      137\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "pulm_severe\n",
      "No          24779\n",
      "Yes          1706\n",
      "Not done      180\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "prim_disease_hct\n",
      "ALL                     8102\n",
      "AML                     7135\n",
      "MDS                     3046\n",
      "IPA                     1719\n",
      "MPN                     1656\n",
      "IEA                     1449\n",
      "NHL                     1319\n",
      "IIS                     1024\n",
      "PCD                      869\n",
      "SAA                      713\n",
      "AI                       449\n",
      "HIS                      445\n",
      "Other leukemia           366\n",
      "Solid tumor              207\n",
      "IMD                      144\n",
      "Other acute leukemia      83\n",
      "HD                        54\n",
      "CML                       20\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "cmv_status\n",
      "+/+    13596\n",
      "-/+     7081\n",
      "+/-     4048\n",
      "-/-     3441\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "tce_imm_match\n",
      "P/P    13114\n",
      "G/G     2522\n",
      "H/H     1084\n",
      "G/B      544\n",
      "H/B      229\n",
      "P/H       83\n",
      "P/B       66\n",
      "P/G       25\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "rituximab\n",
      "No     26033\n",
      "Yes      619\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "prod_type\n",
      "PB    20381\n",
      "BM     8419\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "cyto_score_detail\n",
      "Intermediate    11158\n",
      "Poor             3323\n",
      "Favorable        1208\n",
      "TBD              1043\n",
      "Not tested        145\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "conditioning_intensity\n",
      "MAC                              12288\n",
      "RIC                               7722\n",
      "NMA                               3479\n",
      "TBD                                373\n",
      "No drugs reported                   87\n",
      "N/A, F(pre-TED) not submitted       62\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "ethnicity\n",
      "Not Hispanic or Latino      24482\n",
      "Hispanic or Latino           3347\n",
      "Non-resident of the U.S.      384\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "obesity\n",
      "No          25144\n",
      "Yes          1779\n",
      "Not done      117\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "mrd_hct\n",
      "Negative    8068\n",
      "Positive    4135\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "in_vivo_tcd\n",
      "No     17591\n",
      "Yes    10984\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "tce_match\n",
      "Permissive            6272\n",
      "GvH non-permissive    1605\n",
      "Fully matched         1059\n",
      "HvG non-permissive     868\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "hepatic_severe\n",
      "No          25238\n",
      "Yes          1481\n",
      "Not done      210\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "prior_tumor\n",
      "No          23828\n",
      "Yes          3009\n",
      "Not done      285\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "peptic_ulcer\n",
      "No          25956\n",
      "Yes           259\n",
      "Not done      166\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "gvhd_proph\n",
      "FK+ MMF +- others                  10440\n",
      "Cyclophosphamide alone              5270\n",
      "FK+ MTX +- others(not MMF)          4262\n",
      "Cyclophosphamide +- others          2369\n",
      "CSA + MMF +- others(not FK)         2278\n",
      "FKalone                             1230\n",
      "Other GVHD Prophylaxis               550\n",
      "TDEPLETION alone                     545\n",
      "TDEPLETION +- other                  539\n",
      "No GvHD Prophylaxis                  262\n",
      "CDselect alone                       251\n",
      "CSA + MTX +- others(not MMF,FK)      224\n",
      "CSA alone                            214\n",
      "Parent Q = yes, but no agent          62\n",
      "CDselect +- other                     55\n",
      "CSA +- others(not FK,MMF,MTX)         23\n",
      "FK+- others(not MMF,MTX)               1\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "rheum_issue\n",
      "No          26015\n",
      "Yes           457\n",
      "Not done      145\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "sex_match\n",
      "M-M    7980\n",
      "F-M    7822\n",
      "M-F    6715\n",
      "F-F    6022\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "race_group\n",
      "More than one race                           4845\n",
      "Asian                                        4832\n",
      "White                                        4831\n",
      "Black or African-American                    4795\n",
      "American Indian or Alaska Native             4790\n",
      "Native Hawaiian or other Pacific Islander    4707\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "hepatic_mild\n",
      "No          24989\n",
      "Yes          1754\n",
      "Not done      140\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "tce_div_match\n",
      "Permissive mismatched            12936\n",
      "GvH non-permissive                2458\n",
      "HvG non-permissive                1417\n",
      "Bi-directional non-permissive      593\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "donor_related\n",
      "Related                     16208\n",
      "Unrelated                   12088\n",
      "Multiple donor (non-UCB)      346\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "melphalan_dose\n",
      "N/A, Mel not given    20135\n",
      "MEL                    7260\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "cardiac\n",
      "No          24592\n",
      "Yes          1519\n",
      "Not done      147\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n",
      "pulm_moderate\n",
      "No          21338\n",
      "Yes          5249\n",
      "Not done      166\n",
      "Name: count, dtype: int64\n",
      "----------------------------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "for c in cat:\n",
    "    print(train[c].value_counts())\n",
    "    print('--'*50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "27031d9f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.431109Z",
     "iopub.status.busy": "2025-01-21T13:09:10.430747Z",
     "iopub.status.idle": "2025-01-21T13:09:10.504520Z",
     "shell.execute_reply": "2025-01-21T13:09:10.503784Z"
    },
    "papermill": {
     "duration": 0.09269,
     "end_time": "2025-01-21T13:09:10.506055",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.413365",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "for c in cat:\n",
    "    train[c].fillna('Missing', inplace=True)\n",
    "    test[c].fillna('Missing', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "e11ec1e0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.541707Z",
     "iopub.status.busy": "2025-01-21T13:09:10.541457Z",
     "iopub.status.idle": "2025-01-21T13:09:10.568420Z",
     "shell.execute_reply": "2025-01-21T13:09:10.567514Z"
    },
    "papermill": {
     "duration": 0.044906,
     "end_time": "2025-01-21T13:09:10.569732",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.524826",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def convert_64_to_32(df, num_features):\n",
    "    for c in num_features:\n",
    "        if df[c].dtype == 'float64':\n",
    "            df[c] = df[c].astype('float32')\n",
    "        else:\n",
    "            df[c] = df[c].astype('int32')\n",
    "    return df\n",
    "\n",
    "train = convert_64_to_32(train, num)\n",
    "test = convert_64_to_32(test, num)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "c94a9769",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.621131Z",
     "iopub.status.busy": "2025-01-21T13:09:10.620832Z",
     "iopub.status.idle": "2025-01-21T13:09:10.630617Z",
     "shell.execute_reply": "2025-01-21T13:09:10.630004Z"
    },
    "papermill": {
     "duration": 0.0331,
     "end_time": "2025-01-21T13:09:10.631760",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.598660",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def clean_columns(df):\n",
    "\n",
    "    value_mappings = {\n",
    "        'cmv_status': {\n",
    "            '+/+': 'Positive_Positive',\n",
    "            '+/-': 'Positive_Negative',\n",
    "            '-/+': 'Negative_Positive',\n",
    "            '-/-': 'Negative_Negative'\n",
    "        },\n",
    "        'tbi_status': {\n",
    "            'No TBI': 'No_Total_Body_Irradiation',\n",
    "            'TBI + Cy +- Other': 'TBI_with_Cyclophosphamide_and_Other',\n",
    "            'TBI +- Other, <=cGy': 'TBI_with_Other_Low_Dose',\n",
    "            'TBI +- Other, >cGy': 'TBI_with_Other_High_Dose',\n",
    "            'TBI +- Other, -cGy, single': 'TBI_with_Other_Single_Dose',\n",
    "            'TBI +- Other, unknown dose': 'TBI_with_Other_Unknown_Dose',\n",
    "            'TBI +- Other, -cGy, unknown dose': 'TBI_with_Other_Unknown_Dose',\n",
    "            'TBI +- Other, -cGy, fractionated': 'TBI_with_Other_Fractionated_Dose'\n",
    "        },\n",
    "        'dri_score': {\n",
    "            'Intermediate': 'Intermediate_Risk',\n",
    "            'N/A - pediatric': 'Not_Applicable_Pediatric',\n",
    "            'High': 'High_Risk',\n",
    "            'N/A - non-malignant indication': 'Not_Applicable_Non_Malignant',\n",
    "            'TBD cytogenetics': 'To_Be_Determined_Cytogenetics',\n",
    "            'Low': 'Low_Risk',\n",
    "            'High - TED AML case <missing cytogenetics': 'High_Risk_TED_AML_Missing_Cytogenetics',\n",
    "            'Intermediate - TED AML case <missing cytogenetics': 'Intermediate_Risk_TED_AML_Missing_Cytogenetics',\n",
    "            'N/A - disease not classifiable': 'Not_Applicable_Disease_Not_Classifiable',\n",
    "            'Very high': 'Very_High_Risk',\n",
    "            'Missing disease status': 'Missing'\n",
    "        },\n",
    "        'tce_imm_match': {\n",
    "            'P/P': 'Perfect_Perfect',\n",
    "            'G/G': 'Good_Good',\n",
    "            'H/H': 'Heterozygous_Heterozygous',\n",
    "            'G/B': 'Good_Bad',\n",
    "            'H/B': 'Heterozygous_Bad',\n",
    "            'P/H': 'Perfect_Heterozygous',\n",
    "            'P/B': 'Perfect_Bad',\n",
    "            'P/G': 'Perfect_Good'\n",
    "        },\n",
    "        'gvhd_proph': {\n",
    "            'FK+ MMF +- others': 'Tacrolimus_MMF_with_Others',\n",
    "            'Cyclophosphamide alone': 'Cyclophosphamide_Alone',\n",
    "            'FK+ MTX +- others(not MMF)': 'Tacrolimus_MTX_with_Others_Not_MMF',\n",
    "            'Cyclophosphamide +- others': 'Cyclophosphamide_with_Others',\n",
    "            'CSA + MMF +- others(not FK)': 'Cyclosporine_MMF_with_Others_Not_Tacrolimus',\n",
    "            'FKalone': 'Tacrolimus_Alone',\n",
    "            'Other GVHD Prophylaxis': 'Other_GVHD_Prophylaxis',\n",
    "            'TDEPLETION alone': 'TCell_Depletion_Alone',\n",
    "            'TDEPLETION +- other': 'TCell_Depletion_with_Others',\n",
    "            'No GvHD Prophylaxis': 'No_GVHD_Prophylaxis',\n",
    "            'CDselect alone': 'CD_Selection_Alone',\n",
    "            'CSA + MTX +- others(not MMF,FK)': 'Cyclosporine_MTX_with_Others_Not_MMF_Tacrolimus',\n",
    "            'CSA alone': 'Cyclosporine_Alone',\n",
    "            'Parent Q = yes, but no agent': 'Parent_Yes_No_Agent',\n",
    "            'CDselect +- other': 'CD_Selection_with_Others',\n",
    "            'CSA +- others(not FK,MMF,MTX)': 'Cyclosporine_with_Others_Not_Tacrolimus_MMF_MTX',\n",
    "            'FK+- others(not MMF,MTX)': 'Tacrolimus_with_Others_Not_MMF_MTX'\n",
    "        }\n",
    "    }\n",
    "\n",
    "    for col, mappings in value_mappings.items():\n",
    "        if col in df.columns:\n",
    "            df[col] = df[col].astype(str).map(mappings).fillna(df[col])\n",
    "\n",
    "    return df\n",
    "\n",
    "def clean_space(df):\n",
    "    for col in df.select_dtypes(include=['object']).columns:\n",
    "        df[col] = df[col].apply(lambda x: x.replace(' ', '_') if isinstance(x, str) else x)\n",
    "    return df\n",
    "\n",
    "def clean_not_done(df):\n",
    "    df = df.map(lambda x: 'Missing' if x == 'Not done' else x)\n",
    "    return df\n",
    "\n",
    "def clean_not_tested(df):\n",
    "    df = df.map(lambda x: 'Missing' if x == 'Not tested' else x)\n",
    "    return df\n",
    "\n",
    "def clean_dri_score(score):\n",
    "    if isinstance(score, str) and 'Missing disease status' in score:\n",
    "        return 'Missing'\n",
    "    return score\n",
    "\n",
    "def clean_conditioning_intensity(score):\n",
    "    if isinstance(score, str) and 'No drugs reported' in score:\n",
    "        return 'Missing'\n",
    "    return score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "57b59bfd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:10.665893Z",
     "iopub.status.busy": "2025-01-21T13:09:10.665629Z",
     "iopub.status.idle": "2025-01-21T13:09:11.636653Z",
     "shell.execute_reply": "2025-01-21T13:09:11.635711Z"
    },
    "papermill": {
     "duration": 0.989898,
     "end_time": "2025-01-21T13:09:11.638334",
     "exception": false,
     "start_time": "2025-01-21T13:09:10.648436",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "train = clean_columns(train)\n",
    "test = clean_columns(test)\n",
    "\n",
    "train = clean_not_done(train)\n",
    "test = clean_not_done(test)\n",
    "\n",
    "train = clean_not_tested(train)\n",
    "test = clean_not_tested(test)\n",
    "\n",
    "train['dri_score'] = train['dri_score'].apply(clean_dri_score)\n",
    "test['dri_score'] = test['dri_score'].apply(clean_dri_score)\n",
    "\n",
    "train['conditioning_intensity'] = train['conditioning_intensity'].apply(clean_conditioning_intensity)\n",
    "test['conditioning_intensity'] = test['conditioning_intensity'].apply(clean_conditioning_intensity)\n",
    "\n",
    "train = clean_space(train)\n",
    "test = clean_space(test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d4154c4d",
   "metadata": {
    "papermill": {
     "duration": 0.016212,
     "end_time": "2025-01-21T13:09:11.671966",
     "exception": false,
     "start_time": "2025-01-21T13:09:11.655754",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# LightGBM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "e1cfcd49",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:11.705580Z",
     "iopub.status.busy": "2025-01-21T13:09:11.705297Z",
     "iopub.status.idle": "2025-01-21T13:09:12.136628Z",
     "shell.execute_reply": "2025-01-21T13:09:12.135755Z"
    },
    "papermill": {
     "duration": 0.449625,
     "end_time": "2025-01-21T13:09:12.138193",
     "exception": false,
     "start_time": "2025-01-21T13:09:11.688568",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "\n",
    "def encode_features_lgb(train, test, cat, num):\n",
    "    train = train.copy()\n",
    "    test = test.copy()\n",
    "    \n",
    "    for c in cat:\n",
    "        encoder = LabelEncoder()\n",
    "\n",
    "        # Fit and transform the training data\n",
    "        train[c] = train[c].astype(str)\n",
    "        train[c] = encoder.fit_transform(train[c])\n",
    "        train[c] = train[c].astype('int32').astype('category')\n",
    "\n",
    "        # Transform the test data using the encoder fitted on the training data\n",
    "        test[c] = test[c].astype(str)\n",
    "        test[c] = encoder.transform(test[c])\n",
    "        test[c] = test[c].astype('int32').astype('category')\n",
    "\n",
    "    return train, test\n",
    "\n",
    "train_lgb, test_lgb = encode_features_lgb(train, test, cat, num)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "60806e41",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:12.171568Z",
     "iopub.status.busy": "2025-01-21T13:09:12.171322Z",
     "iopub.status.idle": "2025-01-21T13:09:12.174543Z",
     "shell.execute_reply": "2025-01-21T13:09:12.173923Z"
    },
    "papermill": {
     "duration": 0.021189,
     "end_time": "2025-01-21T13:09:12.176013",
     "exception": false,
     "start_time": "2025-01-21T13:09:12.154824",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "model_params = {\n",
    "    'max_depth': 3,\n",
    "    'colsample_bytree': 0.4,\n",
    "    'learning_rate': 0.02, \n",
    "    'objective': \"regression\", \n",
    "    'verbose': -1, \n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "1d49ff51",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:12.208330Z",
     "iopub.status.busy": "2025-01-21T13:09:12.208105Z",
     "iopub.status.idle": "2025-01-21T13:09:15.741508Z",
     "shell.execute_reply": "2025-01-21T13:09:15.740505Z"
    },
    "papermill": {
     "duration": 3.551761,
     "end_time": "2025-01-21T13:09:15.743400",
     "exception": false,
     "start_time": "2025-01-21T13:09:12.191639",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import StratifiedKFold, train_test_split, KFold\n",
    "from sklearn.metrics import *\n",
    "import lightgbm as lgb\n",
    "from lightgbm import LGBMRegressor\n",
    "\n",
    "def train_lgbm(train, test, model_params, features, cat_features, target):\n",
    "\n",
    "    fix_params = {\n",
    "        'device': \"gpu\",\n",
    "        'cat_features': cat,\n",
    "        'n_estimators': 10000,\n",
    "        'objective': 'regression',\n",
    "        'early_stopping_rounds':20,\n",
    "        'verbose': -1\n",
    "    }\n",
    "    model_params.update(fix_params)\n",
    "    \n",
    "    FOLDS = 10\n",
    "    kf = KFold(n_splits=FOLDS, shuffle=True, random_state=42)\n",
    "    \n",
    "    oof_lgb = np.zeros(len(train))\n",
    "    pred_lgb = np.zeros(len(test))\n",
    "\n",
    "    print('Training model with the followin parameters')\n",
    "    for k, v in model_params.items():\n",
    "        print(f'{k} : {v}')\n",
    "\n",
    "    for i, (t_idx, v_idx) in enumerate(kf.split(train)):\n",
    "\n",
    "        X_train = train.iloc[t_idx][features].copy()\n",
    "        y_train = train.iloc[t_idx][target]\n",
    "        X_valid = train.iloc[v_idx][features].copy()\n",
    "        y_valid = train.iloc[v_idx][target]\n",
    "        X_test = test[features].copy()\n",
    "\n",
    "        model_lgb = LGBMRegressor(**model_params)\n",
    "        model_lgb.fit(\n",
    "            X_train, np.log1p(y_train),\n",
    "            eval_set=[(X_valid, np.log1p(y_valid))],\n",
    "        )  \n",
    "        \n",
    "        y_valid_preds = np.expm1(model_lgb.predict(X_valid))\n",
    "        oof_lgb[v_idx] = y_valid_preds\n",
    "        pred_lgb += np.expm1(model_lgb.predict(X_test))\n",
    "\n",
    "        \n",
    "        fold_rmse = mean_squared_error(y_valid, y_valid_preds, squared=False)\n",
    "        print(\"#\"*25)\n",
    "        print(f\"### Fold {i+1} \\n\")\n",
    "        print(f\"Fold {i+1} RMSE: {fold_rmse}\")\n",
    "        print(\"#\"*25)\n",
    "\n",
    "    pred_lgb /= FOLDS\n",
    "\n",
    "    return model_lgb, oof_lgb, pred_lgb\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "bc6c1622",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:15.780988Z",
     "iopub.status.busy": "2025-01-21T13:09:15.780403Z",
     "iopub.status.idle": "2025-01-21T13:09:15.784041Z",
     "shell.execute_reply": "2025-01-21T13:09:15.783227Z"
    },
    "papermill": {
     "duration": 0.023774,
     "end_time": "2025-01-21T13:09:15.785534",
     "exception": false,
     "start_time": "2025-01-21T13:09:15.761760",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_lgb, oof_lgb, pred_lgb = train_lgbm(\n",
    "#     train=train_lgb, \n",
    "#     test=test_lgb,\n",
    "#     model_params=model_params,\n",
    "#     features=features,\n",
    "#     cat_features=cat,\n",
    "#     target=target\n",
    "# )\n",
    "\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_lgb\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)\n",
    "\n",
    "# # Overall CV for KaplanMeier = 0.672177282559423"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "d2bf5e9b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:15.856485Z",
     "iopub.status.busy": "2025-01-21T13:09:15.856153Z",
     "iopub.status.idle": "2025-01-21T13:09:15.859687Z",
     "shell.execute_reply": "2025-01-21T13:09:15.858874Z"
    },
    "papermill": {
     "duration": 0.02167,
     "end_time": "2025-01-21T13:09:15.860953",
     "exception": false,
     "start_time": "2025-01-21T13:09:15.839283",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# feature_importance = model_lgb.feature_importances_\n",
    "\n",
    "# # Get the feature names from the model (replace with actual feature names if needed)\n",
    "# feature_names = model_lgb.feature_name_\n",
    "\n",
    "# # Sort the feature importances in descending orderA\n",
    "# indices = np.argsort(feature_importance)[::-1]\n",
    "\n",
    "# # Plotting\n",
    "# plt.figure(figsize=(10, 15))  # Increase the figure size for more room\n",
    "# plt.barh(range(len(feature_importance)), feature_importance[indices], align='center')\n",
    "# plt.yticks(range(len(feature_importance)), [feature_names[i] for i in indices], fontsize=10)  # Adjust font size for clarity\n",
    "# plt.xlabel('Feature Importance', fontsize=12)\n",
    "# plt.title('Feature Importance from LightGBM Model', fontsize=10)\n",
    "# plt.gca().invert_yaxis()  # Invert y-axis to show the most important feature at the top\n",
    "\n",
    "# # Adjust layout for padding\n",
    "# plt.tight_layout(pad=1.0)\n",
    "\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4509da43",
   "metadata": {
    "papermill": {
     "duration": 0.01556,
     "end_time": "2025-01-21T13:09:15.892500",
     "exception": false,
     "start_time": "2025-01-21T13:09:15.876940",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# XGBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "fbf9457a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:15.925800Z",
     "iopub.status.busy": "2025-01-21T13:09:15.925522Z",
     "iopub.status.idle": "2025-01-21T13:09:15.931356Z",
     "shell.execute_reply": "2025-01-21T13:09:15.930325Z"
    },
    "papermill": {
     "duration": 0.024075,
     "end_time": "2025-01-21T13:09:15.932572",
     "exception": false,
     "start_time": "2025-01-21T13:09:15.908497",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def encode_features_xgb(train, test, label_encode_features, one_hot_encode_features, features):\n",
    "\n",
    "    # Label Encoding\n",
    "    label_encoders = {}\n",
    "    for feature in label_encode_features:\n",
    "        encoder = LabelEncoder()\n",
    "        train[feature] = train[feature].astype(str)\n",
    "        train[feature] = encoder.fit_transform(train[feature])\n",
    "        test[feature] = test[feature].astype(str)\n",
    "        test[feature] = encoder.transform(test[feature])\n",
    "        label_encoders[feature] = encoder\n",
    "\n",
    "    # One-Hot Encoding\n",
    "    train_one_hot = pd.get_dummies(train[one_hot_encode_features], prefix=one_hot_encode_features)\n",
    "    test_one_hot = pd.get_dummies(test[one_hot_encode_features], prefix=one_hot_encode_features)\n",
    "\n",
    "    train_one_hot, test_one_hot = train_one_hot.align(test_one_hot, join=\"outer\", axis=1, fill_value=0)\n",
    "\n",
    "    train_xgb = pd.concat([train, train_one_hot], axis=1)\n",
    "    test_xgb = pd.concat([test, test_one_hot], axis=1)\n",
    "\n",
    "    features_xgb = features.copy()\n",
    "    features_xgb.extend(train_one_hot.columns)\n",
    "    features_xgb = [f for f in features_xgb if f not in one_hot_encode_features]\n",
    "\n",
    "    return train_xgb, test_xgb, features_xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "ca0d64a0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:15.965105Z",
     "iopub.status.busy": "2025-01-21T13:09:15.964807Z",
     "iopub.status.idle": "2025-01-21T13:09:16.183622Z",
     "shell.execute_reply": "2025-01-21T13:09:16.182559Z"
    },
    "papermill": {
     "duration": 0.236751,
     "end_time": "2025-01-21T13:09:16.185417",
     "exception": false,
     "start_time": "2025-01-21T13:09:15.948666",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Define features for label and one-hot encoding\n",
    "label_encode_features = [\n",
    "    \"dri_score\", \"psych_disturb\", \"cyto_score\", \"diabetes\", \n",
    "    \"arrhythmia\", \"vent_hist\", \"renal_issue\", \"pulm_severe\", \n",
    "    \"cmv_status\", \"tce_imm_match\", \"rituximab\", \"cyto_score_detail\", \n",
    "    \"conditioning_intensity\", \"ethnicity\", \"obesity\", \"mrd_hct\", \n",
    "    \"in_vivo_tcd\", \"tce_match\", \"hepatic_severe\", \"prior_tumor\", \n",
    "    \"peptic_ulcer\", \"gvhd_proph\", \"rheum_issue\", \"sex_match\", \n",
    "    \"hepatic_mild\", \"tce_div_match\", \"donor_related\", \"melphalan_dose\", \n",
    "    \"cardiac\", \"pulm_moderate\"\n",
    "]\n",
    "\n",
    "one_hot_encode_features = [\n",
    "    \"tbi_status\", \"graft_type\", \"prod_type\", \"prim_disease_hct\", \"race_group\"\n",
    "]\n",
    "\n",
    "# Apply the encoding\n",
    "train_xgb, test_xgb, features_xgb = encode_features_xgb(train, test, label_encode_features, one_hot_encode_features, features)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "75f2c537",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:16.220211Z",
     "iopub.status.busy": "2025-01-21T13:09:16.219950Z",
     "iopub.status.idle": "2025-01-21T13:09:16.241637Z",
     "shell.execute_reply": "2025-01-21T13:09:16.240671Z"
    },
    "papermill": {
     "duration": 0.040701,
     "end_time": "2025-01-21T13:09:16.243260",
     "exception": false,
     "start_time": "2025-01-21T13:09:16.202559",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def object_to_cat(df):\n",
    "    for col in df.select_dtypes(include='object').columns:\n",
    "        df[col] = df[col].astype('category')\n",
    "    return df\n",
    "train_xgb2 = object_to_cat(train)\n",
    "test_xgb2 = object_to_cat(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "685b2c7f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:16.278552Z",
     "iopub.status.busy": "2025-01-21T13:09:16.278299Z",
     "iopub.status.idle": "2025-01-21T13:09:16.281857Z",
     "shell.execute_reply": "2025-01-21T13:09:16.281064Z"
    },
    "papermill": {
     "duration": 0.022247,
     "end_time": "2025-01-21T13:09:16.283271",
     "exception": false,
     "start_time": "2025-01-21T13:09:16.261024",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "model_params = {\n",
    "    'device': \"gpu\",\n",
    "    'max_depth': 3,\n",
    "    'colsample_bytree': 0.4,\n",
    "    'learning_rate': 0.02, \n",
    "    'objective': \"regression\", \n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "ae857ff8",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:16.315612Z",
     "iopub.status.busy": "2025-01-21T13:09:16.315345Z",
     "iopub.status.idle": "2025-01-21T13:09:16.551094Z",
     "shell.execute_reply": "2025-01-21T13:09:16.550234Z"
    },
    "papermill": {
     "duration": 0.253282,
     "end_time": "2025-01-21T13:09:16.552574",
     "exception": false,
     "start_time": "2025-01-21T13:09:16.299292",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "from xgboost import XGBRegressor\n",
    "\n",
    "def train_XGB(train, test, model_params, features, target):\n",
    "    fix_params = {\n",
    "        'tree_method': 'hist',\n",
    "        'device': 'cuda', \n",
    "        'objective': 'reg:squarederror',\n",
    "        'n_estimators': 10000,\n",
    "        'early_stopping_rounds': 20,\n",
    "        'eval_metric': 'rmse',\n",
    "        'enable_categorical': True\n",
    "    }\n",
    "    model_params.update(fix_params)\n",
    "\n",
    "    FOLDS = 10\n",
    "    kf = KFold(n_splits=FOLDS, shuffle=True, random_state=42)\n",
    "\n",
    "    oof_xgb = np.zeros(len(train))\n",
    "    pred_xgb = np.zeros(len(test))\n",
    "\n",
    "    print('Training XGBoost model with the following parameters:')\n",
    "    for k, v in model_params.items():\n",
    "        print(f'{k} : {v}')\n",
    "\n",
    "    for i, (t_idx, v_idx) in enumerate(kf.split(train)):\n",
    "        X_train = train.iloc[t_idx][features].copy()\n",
    "        y_train = train.iloc[t_idx][target]\n",
    "        X_valid = train.iloc[v_idx][features].copy()\n",
    "        y_valid = train.iloc[v_idx][target]\n",
    "        X_test = test[features].copy()\n",
    "\n",
    "        model_xgb = XGBRegressor(**model_params)\n",
    "        model_xgb.fit(\n",
    "            X_train, np.log1p(y_train),\n",
    "            eval_set=[(X_valid, np.log1p(y_valid))],\n",
    "            verbose=False\n",
    "        )\n",
    "\n",
    "        y_valid_preds = np.expm1(model_xgb.predict(X_valid))\n",
    "        oof_xgb[v_idx] = y_valid_preds\n",
    "        pred_xgb += np.expm1(model_xgb.predict(X_test))\n",
    "\n",
    "        fold_rmse = mean_squared_error(y_valid, y_valid_preds, squared=False)\n",
    "        print(\"#\" * 25)\n",
    "        print(f\"### Fold {i+1} \\n\")\n",
    "        print(f\"Fold {i+1} RMSE: {fold_rmse}\")\n",
    "        print(\"#\" * 25)\n",
    "\n",
    "    pred_xgb /= FOLDS\n",
    "\n",
    "    return model_xgb, oof_xgb, pred_xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "ec36180f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:16.587558Z",
     "iopub.status.busy": "2025-01-21T13:09:16.587340Z",
     "iopub.status.idle": "2025-01-21T13:09:16.590093Z",
     "shell.execute_reply": "2025-01-21T13:09:16.589426Z"
    },
    "papermill": {
     "duration": 0.021315,
     "end_time": "2025-01-21T13:09:16.591366",
     "exception": false,
     "start_time": "2025-01-21T13:09:16.570051",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_xgb, oof_xgb, pred_xgb = train_XGB(train_xgb, test_xgb, model_params, features_xgb, target)\n",
    "\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_xgb\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "f87c6792",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:16.624797Z",
     "iopub.status.busy": "2025-01-21T13:09:16.624575Z",
     "iopub.status.idle": "2025-01-21T13:09:16.627643Z",
     "shell.execute_reply": "2025-01-21T13:09:16.627047Z"
    },
    "papermill": {
     "duration": 0.02075,
     "end_time": "2025-01-21T13:09:16.628712",
     "exception": false,
     "start_time": "2025-01-21T13:09:16.607962",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_xgb2, oof_xgb2, pred_xgb2 = train_XGB(train_xgb2, test_xgb2, model_params, features, target)\n",
    "\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_xgb2\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8503ab93",
   "metadata": {
    "papermill": {
     "duration": 0.015759,
     "end_time": "2025-01-21T13:09:16.660219",
     "exception": false,
     "start_time": "2025-01-21T13:09:16.644460",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# CatBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "e457e771",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:16.693404Z",
     "iopub.status.busy": "2025-01-21T13:09:16.693190Z",
     "iopub.status.idle": "2025-01-21T13:09:16.706006Z",
     "shell.execute_reply": "2025-01-21T13:09:16.705350Z"
    },
    "papermill": {
     "duration": 0.030789,
     "end_time": "2025-01-21T13:09:16.707408",
     "exception": false,
     "start_time": "2025-01-21T13:09:16.676619",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "train_cat = train.copy()\n",
    "test_cat = test.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "27d7bdae",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:16.742094Z",
     "iopub.status.busy": "2025-01-21T13:09:16.741861Z",
     "iopub.status.idle": "2025-01-21T13:09:16.745028Z",
     "shell.execute_reply": "2025-01-21T13:09:16.744342Z"
    },
    "papermill": {
     "duration": 0.021994,
     "end_time": "2025-01-21T13:09:16.746434",
     "exception": false,
     "start_time": "2025-01-21T13:09:16.724440",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "model_params = {\n",
    "    'depth': 3,\n",
    "    'learning_rate': 0.02, \n",
    "    'loss_function': 'RMSE',\n",
    "    'verbose': -1,  # Suppress verbose output\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "9b8dce83",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:16.778990Z",
     "iopub.status.busy": "2025-01-21T13:09:16.778743Z",
     "iopub.status.idle": "2025-01-21T13:09:17.002669Z",
     "shell.execute_reply": "2025-01-21T13:09:17.001823Z"
    },
    "papermill": {
     "duration": 0.242038,
     "end_time": "2025-01-21T13:09:17.004233",
     "exception": false,
     "start_time": "2025-01-21T13:09:16.762195",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "from catboost import CatBoostRegressor\n",
    "\n",
    "def train_CAT(train, test, model_params, features, cat_features, target):\n",
    "    fix_params = {\n",
    "        'task_type': \"GPU\",\n",
    "        'iterations': 10000,\n",
    "        'loss_function': 'RMSE',\n",
    "        'early_stopping_rounds': 20,\n",
    "    }\n",
    "    model_params.update(fix_params)\n",
    "\n",
    "    FOLDS = 10\n",
    "    kf = KFold(n_splits=FOLDS, shuffle=True, random_state=42)\n",
    "\n",
    "    oof_cat = np.zeros(len(train))\n",
    "    pred_cat = np.zeros(len(test))\n",
    "\n",
    "    print('Training model with the following parameters')\n",
    "    for k, v in model_params.items():\n",
    "        print(f'{k} : {v}')\n",
    "\n",
    "    for i, (t_idx, v_idx) in enumerate(kf.split(train)):\n",
    "        X_train = train.iloc[t_idx][features].copy()\n",
    "        y_train = train.iloc[t_idx][target]\n",
    "        X_valid = train.iloc[v_idx][features].copy()\n",
    "        y_valid = train.iloc[v_idx][target]\n",
    "        X_test = test[features].copy()\n",
    "\n",
    "\n",
    "        model_cat = CatBoostRegressor(**model_params)\n",
    "        model_cat.fit(\n",
    "            X_train, np.log1p(y_train),\n",
    "            cat_features=cat_features,\n",
    "            eval_set=[(X_valid, np.log1p(y_valid))],\n",
    "            verbose=False\n",
    "        )\n",
    "\n",
    "        y_valid_preds = np.expm1(model_cat.predict(X_valid))\n",
    "        oof_cat[v_idx] = y_valid_preds\n",
    "        pred_cat += np.expm1(model_cat.predict(X_test))\n",
    "\n",
    "        fold_rmse = mean_squared_error(y_valid, y_valid_preds, squared=False)\n",
    "        print(\"#\" * 25)\n",
    "        print(f\"### Fold {i + 1} \\n\")\n",
    "        print(f\"Fold {i + 1} RMSE: {fold_rmse}\")\n",
    "        print(\"#\" * 25)\n",
    "\n",
    "    pred_cat /= FOLDS\n",
    "\n",
    "    return model_cat, oof_cat, pred_cat\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "fa559e25",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.038238Z",
     "iopub.status.busy": "2025-01-21T13:09:17.038000Z",
     "iopub.status.idle": "2025-01-21T13:09:17.040743Z",
     "shell.execute_reply": "2025-01-21T13:09:17.040152Z"
    },
    "papermill": {
     "duration": 0.02098,
     "end_time": "2025-01-21T13:09:17.042078",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.021098",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_cat, oof_cat, pred_cat = train_CAT(train_cat, test_cat, model_params, features, cat, target)\n",
    "\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_cat\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "58bb4f6d",
   "metadata": {
    "papermill": {
     "duration": 0.015512,
     "end_time": "2025-01-21T13:09:17.074009",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.058497",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Hyperparameter tuning"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "851f7353",
   "metadata": {
    "papermill": {
     "duration": 0.015686,
     "end_time": "2025-01-21T13:09:17.105529",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.089843",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### For LGBM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "9d7942e5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.138310Z",
     "iopub.status.busy": "2025-01-21T13:09:17.138098Z",
     "iopub.status.idle": "2025-01-21T13:09:17.148351Z",
     "shell.execute_reply": "2025-01-21T13:09:17.147490Z"
    },
    "papermill": {
     "duration": 0.028112,
     "end_time": "2025-01-21T13:09:17.149530",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.121418",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import optuna\n",
    "\n",
    "# Suppress Optuna logs\n",
    "optuna.logging.set_verbosity(optuna.logging.WARNING)\n",
    "\n",
    "\n",
    "def objective_lgbm(trial, data=train_lgb[features], target_lgb=train[target]):\n",
    "    kf = KFold(n_splits=10, shuffle=True, random_state=42)\n",
    "    scores = []\n",
    "\n",
    "    for train_idx, test_idx in kf.split(data):\n",
    "        X_train, X_test = data.iloc[train_idx], data.iloc[test_idx]\n",
    "        y_train, y_test = target.iloc[train_idx], target.iloc[test_idx]\n",
    "\n",
    "\n",
    "        param = {\n",
    "            'device': \"gpu\",\n",
    "            'max_depth': trial.suggest_int('max_depth', 2, 4),\n",
    "            'learning_rate': trial.suggest_float('learning_rate', 1e-3, 0.1, log=True),\n",
    "            'colsample_bytree': trial.suggest_float('colsample_bytree', 0.3, 1.0),\n",
    "            'subsample': trial.suggest_float('subsample', 0.5, 1.0),\n",
    "            'num_leaves': trial.suggest_int('num_leaves', 15, 100),\n",
    "            'min_child_samples': trial.suggest_int('min_child_samples', 20, 50),\n",
    "            'lambda_l1': trial.suggest_float('lambda_l1', 1e-8, 10.0, log=True),\n",
    "            'lambda_l2': trial.suggest_float('lambda_l2', 1e-8, 10.0, log=True),\n",
    "            'bagging_freq': trial.suggest_int('bagging_freq', 1, 10),\n",
    "            'objective': 'regression',\n",
    "            'boosting_type': 'gbdt',\n",
    "            'verbosity': -1,\n",
    "        }\n",
    "\n",
    "        model = LGBMRegressor(**param)\n",
    "        model.fit(\n",
    "            X_train, np.log1p(y_train),\n",
    "            eval_set=[(X_test, np.log1p(y_test))],\n",
    "        )\n",
    "\n",
    "        preds = np.expm1(model.predict(X_test))\n",
    "        rmse = mean_squared_error(y_test, preds, squared=False)\n",
    "        scores.append(rmse)\n",
    "\n",
    "    return np.mean(scores)\n",
    "\n",
    "\n",
    "def progress_callback(study, trial):\n",
    "    print(f\"Trial {trial.number} finished with value: {trial.value}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "8652c13b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.182095Z",
     "iopub.status.busy": "2025-01-21T13:09:17.181840Z",
     "iopub.status.idle": "2025-01-21T13:09:17.184697Z",
     "shell.execute_reply": "2025-01-21T13:09:17.184095Z"
    },
    "papermill": {
     "duration": 0.020665,
     "end_time": "2025-01-21T13:09:17.186033",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.165368",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# study = optuna.create_study(direction='minimize')\n",
    "# study.optimize(objective_lgbm, n_trials=100, callbacks=[progress_callback], n_jobs=-1)\n",
    "# print('Number of finished trials:', len(study.trials))\n",
    "# print('Best trial:', study.best_trial.params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "5e125a1a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.218966Z",
     "iopub.status.busy": "2025-01-21T13:09:17.218708Z",
     "iopub.status.idle": "2025-01-21T13:09:17.221746Z",
     "shell.execute_reply": "2025-01-21T13:09:17.220988Z"
    },
    "papermill": {
     "duration": 0.020973,
     "end_time": "2025-01-21T13:09:17.223097",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.202124",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_params = study.best_trial.params\n",
    "# model_lgb, oof_lgb, pred_lgb = train_lgbm(train_lgb, test_lgb, model_params, features, cat, target)\n",
    "\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_lgb\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "b29eabb2",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.257587Z",
     "iopub.status.busy": "2025-01-21T13:09:17.257371Z",
     "iopub.status.idle": "2025-01-21T13:09:17.261105Z",
     "shell.execute_reply": "2025-01-21T13:09:17.260436Z"
    },
    "papermill": {
     "duration": 0.022418,
     "end_time": "2025-01-21T13:09:17.262193",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.239775",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "model_lgb_params = {\n",
    "    'max_depth': 5,\n",
    "    'learning_rate': 0.01,\n",
    "    'colsample_bytree': 0.40481123886709114,\n",
    "    'subsample': 0.7673666426617842,\n",
    "    'num_leaves': 46,\n",
    "    'min_child_samples': 34,\n",
    "    'lambda_l1': 0.0032893870728708495,\n",
    "    'lambda_l2': 1.5780171816002318e-06,\n",
    "    'bagging_freq': 5,\n",
    "    'cat_features': cat,\n",
    "    'n_estimators': 5000,\n",
    "    'objective': 'regression',\n",
    "    'early_stopping_rounds': 20\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee5ee25c",
   "metadata": {
    "papermill": {
     "duration": 0.016823,
     "end_time": "2025-01-21T13:09:17.295537",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.278714",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### For XGB"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2c1ccf32",
   "metadata": {
    "papermill": {
     "duration": 0.016294,
     "end_time": "2025-01-21T13:09:17.328422",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.312128",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### Manual encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "fb45c7b4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.362371Z",
     "iopub.status.busy": "2025-01-21T13:09:17.362146Z",
     "iopub.status.idle": "2025-01-21T13:09:17.377667Z",
     "shell.execute_reply": "2025-01-21T13:09:17.377094Z"
    },
    "papermill": {
     "duration": 0.034129,
     "end_time": "2025-01-21T13:09:17.378995",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.344866",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def objective_XGB(trial, data=train_xgb[features_xgb], target=train_xgb[target]):\n",
    "    kf = KFold(n_splits=10, shuffle=True, random_state=42)\n",
    "    scores = []\n",
    "\n",
    "    for train_idx, test_idx in kf.split(data):\n",
    "        X_train, X_test = data.iloc[train_idx], data.iloc[test_idx]\n",
    "        y_train, y_test = target.iloc[train_idx], target.iloc[test_idx]\n",
    "\n",
    "        # Define hyperparameters\n",
    "        param = {\n",
    "            'tree_method': 'hist',\n",
    "            'device': 'cuda', \n",
    "            'max_depth': trial.suggest_int('max_depth', 3, 5),\n",
    "            'learning_rate': trial.suggest_float('learning_rate', 1e-3, 0.1, log=True),\n",
    "            'colsample_bytree': trial.suggest_float('colsample_bytree', 0.3, 1.0),\n",
    "            'subsample': trial.suggest_float('subsample', 0.5, 1.0),\n",
    "            'n_estimators': trial.suggest_int('n_estimators', 100, 1000),\n",
    "            'min_child_weight': trial.suggest_float('min_child_weight', 1, 10),\n",
    "            'reg_alpha': trial.suggest_float('reg_alpha', 1e-8, 10.0, log=True),\n",
    "            'reg_lambda': trial.suggest_float('reg_lambda', 1e-8, 10.0, log=True),\n",
    "            'gamma': trial.suggest_float('gamma', 0, 5.0),\n",
    "            'objective': 'reg:squarederror',  \n",
    "            'verbosity': 0,\n",
    "        }\n",
    "\n",
    "\n",
    "        model = XGBRegressor(**param, random_state=42)\n",
    "        model.fit(\n",
    "            X_train, np.log1p(y_train),\n",
    "            eval_set=[(X_test, np.log1p(y_test))],\n",
    "            verbose=False\n",
    "        )\n",
    "\n",
    "\n",
    "        preds = np.expm1(model.predict(X_test))\n",
    "        rmse = mean_squared_error(y_test, preds, squared=False)\n",
    "        scores.append(rmse)\n",
    "\n",
    "    return np.mean(scores)\n",
    "\n",
    "\n",
    "def progress_callback(study, trial):\n",
    "    print(f\"Trial {trial.number} finished with value: {trial.value}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "c6571aa9",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.413011Z",
     "iopub.status.busy": "2025-01-21T13:09:17.412747Z",
     "iopub.status.idle": "2025-01-21T13:09:17.415645Z",
     "shell.execute_reply": "2025-01-21T13:09:17.414960Z"
    },
    "papermill": {
     "duration": 0.021562,
     "end_time": "2025-01-21T13:09:17.416962",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.395400",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# study = optuna.create_study(direction='minimize')\n",
    "# study.optimize(objective_XGB, n_trials=100, callbacks=[progress_callback], n_jobs=-1)\n",
    "# print('Number of finished trials:', len(study.trials))\n",
    "# print('Best trial:', study.best_trial.params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "7f891100",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.451340Z",
     "iopub.status.busy": "2025-01-21T13:09:17.451084Z",
     "iopub.status.idle": "2025-01-21T13:09:17.454393Z",
     "shell.execute_reply": "2025-01-21T13:09:17.453583Z"
    },
    "papermill": {
     "duration": 0.021962,
     "end_time": "2025-01-21T13:09:17.455611",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.433649",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_xgb_params = study.best_trial.params\n",
    "# model_xgb, oof_xgb, pred_xgb = train_XGB(train_lgb, test_lgb, model_xgb_params, features, target)\n",
    "\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_xgb\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)\n",
    "\n",
    "# # Overall CV for KaplanMeier = 0.6715585405549432"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "870a369f",
   "metadata": {
    "papermill": {
     "duration": 0.015918,
     "end_time": "2025-01-21T13:09:17.488186",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.472268",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### Use enable category"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "3c114760",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.520406Z",
     "iopub.status.busy": "2025-01-21T13:09:17.520161Z",
     "iopub.status.idle": "2025-01-21T13:09:17.534814Z",
     "shell.execute_reply": "2025-01-21T13:09:17.533973Z"
    },
    "papermill": {
     "duration": 0.032447,
     "end_time": "2025-01-21T13:09:17.536324",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.503877",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def objective_XGB2(trial, data=train_xgb2[features], target=train_xgb2[target]):\n",
    "    kf = KFold(n_splits=10, shuffle=True, random_state=42)\n",
    "    scores = []\n",
    "\n",
    "    for train_idx, test_idx in kf.split(data):\n",
    "        X_train, X_test = data.iloc[train_idx], data.iloc[test_idx]\n",
    "        y_train, y_test = target.iloc[train_idx], target.iloc[test_idx]\n",
    "\n",
    "        # Define hyperparameters\n",
    "        param = {\n",
    "            'tree_method': 'gpu_hist',\n",
    "            'enable_categorical': True,\n",
    "            'max_depth': trial.suggest_int('max_depth', 3, 5),\n",
    "            'learning_rate': trial.suggest_float('learning_rate', 1e-3, 0.1, log=True),\n",
    "            'colsample_bytree': trial.suggest_float('colsample_bytree', 0.3, 1.0),\n",
    "            'subsample': trial.suggest_float('subsample', 0.5, 1.0),\n",
    "            'n_estimators': trial.suggest_int('n_estimators', 100, 1000),\n",
    "            'min_child_weight': trial.suggest_float('min_child_weight', 1, 10),\n",
    "            'reg_alpha': trial.suggest_float('reg_alpha', 1e-8, 10.0, log=True),\n",
    "            'reg_lambda': trial.suggest_float('reg_lambda', 1e-8, 10.0, log=True),\n",
    "            'gamma': trial.suggest_float('gamma', 0, 5.0),\n",
    "            'objective': 'reg:squarederror',  \n",
    "            'verbosity': 0,\n",
    "        }\n",
    "\n",
    "\n",
    "        model = XGBRegressor(**param, random_state=42)\n",
    "        model.fit(\n",
    "            X_train, np.log1p(y_train),\n",
    "            eval_set=[(X_test, np.log1p(y_test))],\n",
    "            verbose=False\n",
    "        )\n",
    "\n",
    "\n",
    "        preds = np.expm1(model.predict(X_test))\n",
    "        rmse = mean_squared_error(y_test, preds, squared=False)\n",
    "        scores.append(rmse)\n",
    "\n",
    "    return np.mean(scores)\n",
    "\n",
    "\n",
    "def progress_callback(study, trial):\n",
    "    print(f\"Trial {trial.number} finished with value: {trial.value}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "f789f76e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.570150Z",
     "iopub.status.busy": "2025-01-21T13:09:17.569830Z",
     "iopub.status.idle": "2025-01-21T13:09:17.572713Z",
     "shell.execute_reply": "2025-01-21T13:09:17.572001Z"
    },
    "papermill": {
     "duration": 0.021222,
     "end_time": "2025-01-21T13:09:17.574057",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.552835",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# study = optuna.create_study(direction='minimize')\n",
    "# study.optimize(objective_XGB2, n_trials=100, callbacks=[progress_callback], n_jobs=-1)\n",
    "# print('Number of finished trials:', len(study.trials))\n",
    "# print('Best trial:', study.best_trial.params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "ad6be56a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.607943Z",
     "iopub.status.busy": "2025-01-21T13:09:17.607698Z",
     "iopub.status.idle": "2025-01-21T13:09:17.610384Z",
     "shell.execute_reply": "2025-01-21T13:09:17.609789Z"
    },
    "papermill": {
     "duration": 0.021057,
     "end_time": "2025-01-21T13:09:17.611666",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.590609",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_xgb_params = study.best_trial.params\n",
    "# model_xgb, oof_xgb, pred_xgb = train_XGB(train_lgb, test_lgb, model_xgb_params, features, target)\n",
    "\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_xgb\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "9e43e6fb",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.645174Z",
     "iopub.status.busy": "2025-01-21T13:09:17.644879Z",
     "iopub.status.idle": "2025-01-21T13:09:17.649232Z",
     "shell.execute_reply": "2025-01-21T13:09:17.648614Z"
    },
    "papermill": {
     "duration": 0.022202,
     "end_time": "2025-01-21T13:09:17.650434",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.628232",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "model_xgb_params = {\n",
    "    'max_depth': 4, \n",
    "    'learning_rate': 0.09180807102095336,\n",
    "    'colsample_bytree': 0.3809438487844099, \n",
    "    'subsample': 0.844622438351228, \n",
    "    'n_estimators': 419, \n",
    "    'min_child_weight': 3.714743419003562, \n",
    "    'reg_alpha': 5.80197653137552e-06, \n",
    "    'reg_lambda': 1.2374095115455325e-08, \n",
    "    'gamma': 0.0037460722016019465\n",
    "}\n",
    "\n",
    "model_xgb2_params = {\n",
    "    \"max_depth\": 9,\n",
    "    \"learning_rate\": 0.018203874021653552,\n",
    "    \"colsample_bytree\": 0.41392312362600636,\n",
    "    \"subsample\": 0.870771567534879,\n",
    "    \"n_estimators\": 10000,\n",
    "    \"min_child_weight\": 6.587958958652532,\n",
    "    \"reg_alpha\": 1.675358492618636e-07,\n",
    "    \"reg_lambda\": 0.004228750471811781,\n",
    "    \"gamma\": 0.02009243264106564,\n",
    "    \"tree_method\": \"gpu_hist\",\n",
    "    \"objective\": \"reg:squarederror\",\n",
    "    \"early_stopping_rounds\": 20,\n",
    "    \"eval_metric\": \"rmse\",\n",
    "    \"enable_categorical\": True\n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "1e7b25cd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.683411Z",
     "iopub.status.busy": "2025-01-21T13:09:17.683174Z",
     "iopub.status.idle": "2025-01-21T13:09:17.686206Z",
     "shell.execute_reply": "2025-01-21T13:09:17.685343Z"
    },
    "papermill": {
     "duration": 0.021099,
     "end_time": "2025-01-21T13:09:17.687604",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.666505",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_params = study.best_trial.params\n",
    "# model_xgb, oof_xgb, pred_xgb = train_XGB(train_lgb, test_lgb, model_xgb_params, features, target)\n",
    "\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_xgb\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "db691595",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.721515Z",
     "iopub.status.busy": "2025-01-21T13:09:17.721236Z",
     "iopub.status.idle": "2025-01-21T13:09:17.724338Z",
     "shell.execute_reply": "2025-01-21T13:09:17.723535Z"
    },
    "papermill": {
     "duration": 0.021381,
     "end_time": "2025-01-21T13:09:17.725638",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.704257",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_xgb, oof_xgb, pred_xgb = train_XGB(train_xgb, test_xgb, model_xgb_params, features_xgb, target)\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_xgb\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)\n",
    "\n",
    "# # XGB 2 is better"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84c72e97",
   "metadata": {
    "papermill": {
     "duration": 0.016432,
     "end_time": "2025-01-21T13:09:17.758464",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.742032",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### For CAT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "39ce0292",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.791381Z",
     "iopub.status.busy": "2025-01-21T13:09:17.791119Z",
     "iopub.status.idle": "2025-01-21T13:09:17.811106Z",
     "shell.execute_reply": "2025-01-21T13:09:17.810149Z"
    },
    "papermill": {
     "duration": 0.038337,
     "end_time": "2025-01-21T13:09:17.812590",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.774253",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def objective_CAT(trial, data=train[features], target=train[target], cat_features=cat):\n",
    "    kf = KFold(n_splits=3, shuffle=True, random_state=42)\n",
    "    scores = []\n",
    "\n",
    "    for train_idx, test_idx in kf.split(data):\n",
    "        X_train, X_test = data.iloc[train_idx], data.iloc[test_idx]\n",
    "        y_train, y_test = target.iloc[train_idx], target.iloc[test_idx]\n",
    "\n",
    "        param = {\n",
    "            'task_type': \"GPU\",\n",
    "            'loss_function': 'RMSE',\n",
    "            'depth': trial.suggest_int('depth', 3, 10),\n",
    "            'learning_rate': trial.suggest_float('learning_rate', 1e-3, 0.1, log=True),\n",
    "            'l2_leaf_reg': trial.suggest_float('l2_leaf_reg', 1e-8, 10.0, log=True),\n",
    "            'random_strength': trial.suggest_float('random_strength', 0.0, 10.0),\n",
    "            'bootstrap_type': trial.suggest_categorical('bootstrap_type', ['Bayesian', 'Bernoulli', 'MVS']),\n",
    "            'verbose': 0\n",
    "        }\n",
    "\n",
    "        # Conditionally add parameters based on bootstrap type\n",
    "        if param['bootstrap_type'] == 'Bayesian':\n",
    "            param['bagging_temperature'] = trial.suggest_float('bagging_temperature', 0.0, 10.0)\n",
    "        elif param['bootstrap_type'] == 'Bernoulli':\n",
    "            param['subsample'] = trial.suggest_float('subsample', 0.5, 1.0)\n",
    "\n",
    "        model_cat = CatBoostRegressor(**param)\n",
    "        model_cat.fit(\n",
    "            X_train, np.log1p(y_train),\n",
    "            cat_features=cat_features,\n",
    "            eval_set=[(X_test, np.log1p(y_test))],\n",
    "            verbose=False,\n",
    "            early_stopping_rounds=20\n",
    "        )\n",
    "\n",
    "        preds = np.expm1(model_cat.predict(X_test))\n",
    "        rmse = mean_squared_error(y_test, preds, squared=False)\n",
    "        scores.append(rmse)\n",
    "\n",
    "    return np.mean(scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "06d7c823",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.846206Z",
     "iopub.status.busy": "2025-01-21T13:09:17.845965Z",
     "iopub.status.idle": "2025-01-21T13:09:17.848846Z",
     "shell.execute_reply": "2025-01-21T13:09:17.848235Z"
    },
    "papermill": {
     "duration": 0.020224,
     "end_time": "2025-01-21T13:09:17.849931",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.829707",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# study = optuna.create_study(direction='minimize')\n",
    "# study.optimize(objective_CAT, n_trials=20, callbacks=[progress_callback],  n_jobs=1)\n",
    "# print('Number of finished trials:', len(study.trials))\n",
    "# print('Best trial:', study.best_trial.params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "d8af20ce",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.882370Z",
     "iopub.status.busy": "2025-01-21T13:09:17.882158Z",
     "iopub.status.idle": "2025-01-21T13:09:17.885234Z",
     "shell.execute_reply": "2025-01-21T13:09:17.884479Z"
    },
    "papermill": {
     "duration": 0.020732,
     "end_time": "2025-01-21T13:09:17.886527",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.865795",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_params = study.best_trial.params\n",
    "# model_cat, oof_cat, pred_cat = train_CAT(train_cat, test_cat, model_params, features, cat, target)\n",
    "\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_cat\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "abbd5fe3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.919115Z",
     "iopub.status.busy": "2025-01-21T13:09:17.918914Z",
     "iopub.status.idle": "2025-01-21T13:09:17.921671Z",
     "shell.execute_reply": "2025-01-21T13:09:17.921131Z"
    },
    "papermill": {
     "duration": 0.020792,
     "end_time": "2025-01-21T13:09:17.922922",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.902130",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# model_params = {\n",
    "#     'depth': 5, \n",
    "#     'learning_rate': 0.05259516359861675, \n",
    "#     'bagging_temperature':0.5,\n",
    "#     'l2_leaf_reg': 6.806823646372654e-06, \n",
    "#     'random_strength': 5.889614035287661\n",
    "# }\n",
    "\n",
    "# model_cat, oof_cat, pred_cat = train_CAT(train_cat, test_cat, model_params, features, cat, target)\n",
    "\n",
    "# y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "# y_pred = train[[\"ID\"]].copy()\n",
    "# y_pred[\"prediction\"] = oof_cat\n",
    "# m = score(y_true, y_pred, \"ID\")\n",
    "# print(f\"\\nOverall CV for KaplanMeier =\", m)\n",
    "\n",
    "# # Overall CV for KaplanMeier = 0.6697462480992807 (depth = 9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "9da2dd59",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:17.956512Z",
     "iopub.status.busy": "2025-01-21T13:09:17.956270Z",
     "iopub.status.idle": "2025-01-21T13:09:17.959008Z",
     "shell.execute_reply": "2025-01-21T13:09:17.958400Z"
    },
    "papermill": {
     "duration": 0.020605,
     "end_time": "2025-01-21T13:09:17.960259",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.939654",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Submission"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b129878",
   "metadata": {
    "papermill": {
     "duration": 0.015737,
     "end_time": "2025-01-21T13:09:17.991625",
     "exception": false,
     "start_time": "2025-01-21T13:09:17.975888",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Ensemble"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "3e8385ad",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:09:18.026340Z",
     "iopub.status.busy": "2025-01-21T13:09:18.026109Z",
     "iopub.status.idle": "2025-01-21T13:11:50.207825Z",
     "shell.execute_reply": "2025-01-21T13:11:50.207122Z"
    },
    "papermill": {
     "duration": 152.200312,
     "end_time": "2025-01-21T13:11:50.209033",
     "exception": false,
     "start_time": "2025-01-21T13:09:18.008721",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training model with the followin parameters\n",
      "max_depth : 5\n",
      "learning_rate : 0.01\n",
      "colsample_bytree : 0.40481123886709114\n",
      "subsample : 0.7673666426617842\n",
      "num_leaves : 46\n",
      "min_child_samples : 34\n",
      "lambda_l1 : 0.0032893870728708495\n",
      "lambda_l2 : 1.5780171816002318e-06\n",
      "bagging_freq : 5\n",
      "cat_features : ['dri_score', 'psych_disturb', 'cyto_score', 'diabetes', 'tbi_status', 'arrhythmia', 'graft_type', 'vent_hist', 'renal_issue', 'pulm_severe', 'prim_disease_hct', 'cmv_status', 'tce_imm_match', 'rituximab', 'prod_type', 'cyto_score_detail', 'conditioning_intensity', 'ethnicity', 'obesity', 'mrd_hct', 'in_vivo_tcd', 'tce_match', 'hepatic_severe', 'prior_tumor', 'peptic_ulcer', 'gvhd_proph', 'rheum_issue', 'sex_match', 'race_group', 'hepatic_mild', 'tce_div_match', 'donor_related', 'melphalan_dose', 'cardiac', 'pulm_moderate']\n",
      "n_estimators : 10000\n",
      "objective : regression\n",
      "early_stopping_rounds : 20\n",
      "device : gpu\n",
      "verbose : -1\n",
      "#########################\n",
      "### Fold 1 \n",
      "\n",
      "Fold 1 RMSE: 0.15663867881999133\n",
      "#########################\n",
      "#########################\n",
      "### Fold 2 \n",
      "\n",
      "Fold 2 RMSE: 0.15324116212840877\n",
      "#########################\n",
      "#########################\n",
      "### Fold 3 \n",
      "\n",
      "Fold 3 RMSE: 0.15529344003462048\n",
      "#########################\n",
      "#########################\n",
      "### Fold 4 \n",
      "\n",
      "Fold 4 RMSE: 0.15819909004525495\n",
      "#########################\n",
      "#########################\n",
      "### Fold 5 \n",
      "\n",
      "Fold 5 RMSE: 0.1540413566904352\n",
      "#########################\n",
      "#########################\n",
      "### Fold 6 \n",
      "\n",
      "Fold 6 RMSE: 0.15764355278070913\n",
      "#########################\n",
      "#########################\n",
      "### Fold 7 \n",
      "\n",
      "Fold 7 RMSE: 0.15916146780771176\n",
      "#########################\n",
      "#########################\n",
      "### Fold 8 \n",
      "\n",
      "Fold 8 RMSE: 0.15438764737916452\n",
      "#########################\n",
      "#########################\n",
      "### Fold 9 \n",
      "\n",
      "Fold 9 RMSE: 0.157418211240835\n",
      "#########################\n",
      "#########################\n",
      "### Fold 10 \n",
      "\n",
      "Fold 10 RMSE: 0.15545175760093827\n",
      "#########################\n",
      "Training XGBoost model with the following parameters:\n",
      "max_depth : 4\n",
      "learning_rate : 0.09180807102095336\n",
      "colsample_bytree : 0.3809438487844099\n",
      "subsample : 0.844622438351228\n",
      "n_estimators : 10000\n",
      "min_child_weight : 3.714743419003562\n",
      "reg_alpha : 5.80197653137552e-06\n",
      "reg_lambda : 1.2374095115455325e-08\n",
      "gamma : 0.0037460722016019465\n",
      "tree_method : hist\n",
      "device : cuda\n",
      "objective : reg:squarederror\n",
      "early_stopping_rounds : 20\n",
      "eval_metric : rmse\n",
      "enable_categorical : True\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.10/dist-packages/xgboost/core.py:160: UserWarning: [13:10:59] WARNING: /workspace/src/common/error_msg.cc:58: Falling back to prediction using DMatrix due to mismatched devices. This might lead to higher memory usage and slower performance. XGBoost is running on: cuda:0, while the input data is on: cpu.\n",
      "Potential solutions:\n",
      "- Use a data structure that matches the device ordinal in the booster.\n",
      "- Set the device for booster before call to inplace_predict.\n",
      "\n",
      "This warning will only be shown once.\n",
      "\n",
      "  warnings.warn(smsg, UserWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#########################\n",
      "### Fold 1 \n",
      "\n",
      "Fold 1 RMSE: 0.1570957662784397\n",
      "#########################\n",
      "#########################\n",
      "### Fold 2 \n",
      "\n",
      "Fold 2 RMSE: 0.15407803661045683\n",
      "#########################\n",
      "#########################\n",
      "### Fold 3 \n",
      "\n",
      "Fold 3 RMSE: 0.15644834819681955\n",
      "#########################\n",
      "#########################\n",
      "### Fold 4 \n",
      "\n",
      "Fold 4 RMSE: 0.1582960988259115\n",
      "#########################\n",
      "#########################\n",
      "### Fold 5 \n",
      "\n",
      "Fold 5 RMSE: 0.15546561482922305\n",
      "#########################\n",
      "#########################\n",
      "### Fold 6 \n",
      "\n",
      "Fold 6 RMSE: 0.15874251625850785\n",
      "#########################\n",
      "#########################\n",
      "### Fold 7 \n",
      "\n",
      "Fold 7 RMSE: 0.15920835950420714\n",
      "#########################\n",
      "#########################\n",
      "### Fold 8 \n",
      "\n",
      "Fold 8 RMSE: 0.15521831677229558\n",
      "#########################\n",
      "#########################\n",
      "### Fold 9 \n",
      "\n",
      "Fold 9 RMSE: 0.15831704494907278\n",
      "#########################\n",
      "#########################\n",
      "### Fold 10 \n",
      "\n",
      "Fold 10 RMSE: 0.15575190654713347\n",
      "#########################\n",
      "Training XGBoost model with the following parameters:\n",
      "max_depth : 9\n",
      "learning_rate : 0.018203874021653552\n",
      "colsample_bytree : 0.41392312362600636\n",
      "subsample : 0.870771567534879\n",
      "n_estimators : 10000\n",
      "min_child_weight : 6.587958958652532\n",
      "reg_alpha : 1.675358492618636e-07\n",
      "reg_lambda : 0.004228750471811781\n",
      "gamma : 0.02009243264106564\n",
      "tree_method : hist\n",
      "objective : reg:squarederror\n",
      "early_stopping_rounds : 20\n",
      "eval_metric : rmse\n",
      "enable_categorical : True\n",
      "device : cuda\n",
      "#########################\n",
      "### Fold 1 \n",
      "\n",
      "Fold 1 RMSE: 0.15790880962077303\n",
      "#########################\n",
      "#########################\n",
      "### Fold 2 \n",
      "\n",
      "Fold 2 RMSE: 0.1537923303369023\n",
      "#########################\n",
      "#########################\n",
      "### Fold 3 \n",
      "\n",
      "Fold 3 RMSE: 0.155866218604957\n",
      "#########################\n",
      "#########################\n",
      "### Fold 4 \n",
      "\n",
      "Fold 4 RMSE: 0.15812157239076532\n",
      "#########################\n",
      "#########################\n",
      "### Fold 5 \n",
      "\n",
      "Fold 5 RMSE: 0.15509919226032148\n",
      "#########################\n",
      "#########################\n",
      "### Fold 6 \n",
      "\n",
      "Fold 6 RMSE: 0.1586041056066703\n",
      "#########################\n",
      "#########################\n",
      "### Fold 7 \n",
      "\n",
      "Fold 7 RMSE: 0.16004529658485062\n",
      "#########################\n",
      "#########################\n",
      "### Fold 8 \n",
      "\n",
      "Fold 8 RMSE: 0.1549884201476021\n",
      "#########################\n",
      "#########################\n",
      "### Fold 9 \n",
      "\n",
      "Fold 9 RMSE: 0.15855273287390134\n",
      "#########################\n",
      "#########################\n",
      "### Fold 10 \n",
      "\n",
      "Fold 10 RMSE: 0.1562045979580824\n",
      "#########################\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-2-a1aa375cd29b>:53: FutureWarning: The default of observed=False is deprecated and will be changed to True in a future version of pandas. Pass observed=False to retain current behavior or observed=True to adopt the future default and silence this warning.\n",
      "  merged_df_race_dict = dict(merged_df.groupby(['race_group']).groups)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Overall CV for KaplanMeier = 0.6745155633466419\n"
     ]
    }
   ],
   "source": [
    "model_lgb_params = {\n",
    "    'max_depth': 5,\n",
    "    'learning_rate': 0.01,\n",
    "    'colsample_bytree': 0.40481123886709114,\n",
    "    'subsample': 0.7673666426617842,\n",
    "    'num_leaves': 46,\n",
    "    'min_child_samples': 34,\n",
    "    'lambda_l1': 0.0032893870728708495,\n",
    "    'lambda_l2': 1.5780171816002318e-06,\n",
    "    'bagging_freq': 5,\n",
    "    'cat_features': cat,\n",
    "    'n_estimators': 5000,\n",
    "    'objective': 'regression',\n",
    "    'early_stopping_rounds': 20\n",
    "}\n",
    "\n",
    "model_xgb_params = {\n",
    "    'max_depth': 4, \n",
    "    'learning_rate': 0.09180807102095336,\n",
    "    'colsample_bytree': 0.3809438487844099, \n",
    "    'subsample': 0.844622438351228, \n",
    "    'n_estimators': 419, \n",
    "    'min_child_weight': 3.714743419003562, \n",
    "    'reg_alpha': 5.80197653137552e-06, \n",
    "    'reg_lambda': 1.2374095115455325e-08, \n",
    "    'gamma': 0.0037460722016019465\n",
    "}\n",
    "\n",
    "model_xgb2_params = {\n",
    "    \"max_depth\": 9,\n",
    "    \"learning_rate\": 0.018203874021653552,\n",
    "    \"colsample_bytree\": 0.41392312362600636,\n",
    "    \"subsample\": 0.870771567534879,\n",
    "    \"n_estimators\": 10000,\n",
    "    \"min_child_weight\": 6.587958958652532,\n",
    "    \"reg_alpha\": 1.675358492618636e-07,\n",
    "    \"reg_lambda\": 0.004228750471811781,\n",
    "    \"gamma\": 0.02009243264106564,\n",
    "    \"tree_method\": \"gpu_hist\",\n",
    "    \"objective\": \"reg:squarederror\",\n",
    "    \"early_stopping_rounds\": 20,\n",
    "    \"eval_metric\": \"rmse\",\n",
    "    \"enable_categorical\": True\n",
    "}\n",
    "\n",
    "\n",
    "# model_cat_params = {\n",
    "#     'depth': 5, \n",
    "#     'learning_rate': 0.05259516359861675, \n",
    "#     'bagging_temperature':0.5,\n",
    "#     'l2_leaf_reg': 6.806823646372654e-06, \n",
    "#     'random_strength': 5.889614035287661\n",
    "# }\n",
    "\n",
    "model_lgb, oof_lgb, pred_lgb = train_lgbm(train_lgb, test_lgb, model_lgb_params, features, cat, target)\n",
    "model_xgb, oof_xgb, pred_xgb = train_XGB(train_xgb, test_xgb, model_xgb_params, features_xgb, target)\n",
    "model_xgb2, oof_xgb2, pred_xgb2 = train_XGB(train_xgb2, test_xgb2, model_xgb2_params, features, target)\n",
    "\n",
    "oof_ensemble = (oof_lgb + oof_xgb + oof_xgb2) / 3\n",
    "pred_ensemble = (pred_lgb + pred_xgb + pred_xgb2) / 3\n",
    "\n",
    "y_true = train[[\"ID\",\"efs\",\"efs_time\",\"race_group\"]].copy()\n",
    "y_pred = train[[\"ID\"]].copy()\n",
    "y_pred[\"prediction\"] = oof_ensemble\n",
    "m = score(y_true, y_pred, \"ID\")\n",
    "print(f\"\\nOverall CV for KaplanMeier =\", m)\n",
    "# Overall CV for KaplanMeier = 0.6744979717032406 (with 2 xgb)\n",
    "# Overall CV for KaplanMeier = 0.6742222404834405 (with 1 xgb)\n",
    "# Overall CV for KaplanMeier = 0.6745101916302204 (with 2 xgb, avg preds)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8a70b2c6",
   "metadata": {
    "papermill": {
     "duration": 0.019591,
     "end_time": "2025-01-21T13:11:50.248818",
     "exception": false,
     "start_time": "2025-01-21T13:11:50.229227",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "14f3430d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-01-21T13:11:50.287062Z",
     "iopub.status.busy": "2025-01-21T13:11:50.286773Z",
     "iopub.status.idle": "2025-01-21T13:11:50.302403Z",
     "shell.execute_reply": "2025-01-21T13:11:50.301505Z"
    },
    "papermill": {
     "duration": 0.035037,
     "end_time": "2025-01-21T13:11:50.303591",
     "exception": false,
     "start_time": "2025-01-21T13:11:50.268554",
     "status": "completed"
    },
    "tags": []
   },
   "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>ID</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>28800</td>\n",
       "      <td>0.506241</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>28801</td>\n",
       "      <td>0.633764</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28802</td>\n",
       "      <td>0.465576</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      ID  prediction\n",
       "0  28800    0.506241\n",
       "1  28801    0.633764\n",
       "2  28802    0.465576"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_submission = pd.read_csv('/kaggle/input/equity-post-HCT-survival-predictions/sample_submission.csv')\n",
    "submission = sample_submission.copy()\n",
    "submission['prediction'] = pred_ensemble\n",
    "submission.to_csv('/kaggle/working/submission.csv', index=False)\n",
    "submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2925e228",
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    },
    "papermill": {
     "duration": 0.017247,
     "end_time": "2025-01-21T13:11:50.339104",
     "exception": false,
     "start_time": "2025-01-21T13:11:50.321857",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kaggle": {
   "accelerator": "nvidiaTeslaT4",
   "dataSources": [
    {
     "databundleVersionId": 10381525,
     "sourceId": 70942,
     "sourceType": "competition"
    },
    {
     "sourceId": 211253469,
     "sourceType": "kernelVersion"
    },
    {
     "sourceId": 211322530,
     "sourceType": "kernelVersion"
    }
   ],
   "dockerImageVersionId": 30822,
   "isGpuEnabled": true,
   "isInternetEnabled": false,
   "language": "python",
   "sourceType": "notebook"
  },
  "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": "3.10.12"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 188.403445,
   "end_time": "2025-01-21T13:11:51.476013",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2025-01-21T13:08:43.072568",
   "version": "2.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}