Project information

  • Category: Predictive Analysis, Neural Network
  • Reason: Purdue Health Tech Challenge
  • Project date: 2025
  • Project URL: Click Here

More about this project:

A PyTorch-based feedforward neural network for predicting cancer severity levels from medical tabular data, with automatic patient risk stratification using K-means clustering.

This project uses deep learning to predict cancer severity scores (0-1 scale) from patient medical records. The model is trained on NIH Glioblastoma dataset containing 21,634 patient records with 164 clinical features.