Upload GVHD_Final_Report.ipynb

GVHD_Final_Report.ipynb

@@ -0,0 +1,92 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# GVHD Severity Prediction - Final Report\n",
+    "\n",
+    "**GPU-Optimized Pipeline with Feature Engineering and Stacking**\n",
+    "\n",
+    "| Metric | Value |\n",
+    "|--------|-------|\n",
+    "| Best Model | Stacking Ensemble |\n",
+    "| **AUC** | **0.7083 ± 0.0117** |\n",
+    "| Baseline AUC | 0.7034 |\n",
+    "| Improvement | +0.0049 (+0.7%) |\n",
+    "| Brier (Platt) | 0.2019 |\n",
+    "| Threshold | 0.544 |\n",
+    "| Sensitivity | 68.4% |\n",
+    "| Specificity | 58.8% |\n",
+    "| PPV | 76.9% |\n",
+    "| NPV | 48.3% |"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Table 1: Model Comparison\n",
+    "\n",
+    "| Model | AUC Mean | AUC Std | Fold1 | Fold2 | Fold3 | Fold4 | Fold5 |\n",
+    "|-------|----------|---------|-------|-------|-------|-------|-------|\n",
+    "| CatBoost | 0.6963 | ±0.0105 | 0.700 | 0.711 | 0.699 | 0.679 | 0.694 |\n",
+    "| XGBoost | 0.6986 | ±0.0126 | 0.705 | 0.711 | 0.704 | 0.675 | 0.698 |\n",
+    "| NeuralNet | 0.6870 | ±0.0088 | 0.692 | 0.699 | 0.698 | 0.677 | 0.681 |\n",
+    "| **Stacking** | **0.7083** | **±0.0117** | **0.714** | **0.722** | **0.714** | **0.688** | **0.703** |"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Table 2: Calibration Assessment\n",
+    "\n",
+    "| Method | Brier Score |\n",
+    "|--------|------------|\n",
+    "| Raw | 0.2150 |\n",
+    "| Platt Scaling | 0.2019 |\n",
+    "| Isotonic | 0.2024 |"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Table 3: Optimal Threshold\n",
+    "\n",
+    "| Metric | Value |\n",
+    "|--------|-------|\n",
+    "| Youden's J | 0.544 |\n",
+    "| Sensitivity | 68.4% |\n",
+    "| Specificity | 58.8% |\n",
+    "| PPV | 76.9% |\n",
+    "| NPV | 48.3% |"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "results = pd.read_csv('/app/result_comparison_final.csv')\n",
+    "print(results.to_string(index=False))"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "name": "python",
+   "version": "3.12.0"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}