1. Overview

This project tests DeepHit, a discrete‐time competing‐risks survival model, implemented using PyTorch and Pycox. The pipeline consists of the following steps:

• Data loading & splitting
  Stratified train/validation/test splits that preserve event‐rate distributions.

• Preprocessing
  Feature standardization and discrete‐time label transformation (NUM_DURATIONS = 16, equidistant intervals).

• Model
  A simple 2‐layer MLP predicting per‐event, per‐time‐step risks.

• Training
  Uses Adam optimizer with ranking + likelihood loss (from DeepHit), and early stopping on validation loss.

• Evaluation

  • Time‐dependent concordance (Antolini)
  • IPCW C-index at quartile time points
  • Brier scores for each event at specified horizons

---

2. Usage Instructions

2.1. Install dependencies

pip install -r requirements.txt

2.2. Run training and evaluation

python src/main.py --data ./data/SYNTHETIC/synthetic_comprisk.csv --durations 16 --batch-size 128 --epochs 100 --lr 0.005

2.3. Inspect results and plots

• Loss curves and CIF plots are saved to outputs/figures/
• Evaluation metrics are logged in outputs/metrics.json

---

3. Edge Cases

• Single Event (No Competing Risk)
  To test with only one critical event:

  python src/main.py --data ./data/SYNTHETIC/synthetic_one_label.csv --durations 16 --batch-size 128 --epochs 100 --lr 0.005

• All Censored Observations
  If all observations are censored, there is no direct information on event-time distributions, only that events exceed the last follow-up.
  In this case, neither nonparametric nor semiparametric methods can estimate survival beyond confirming it remains at 1 until the last censoring time.