README.md

metadata

title: CVD Risk Predictor
sdk: streamlit
emoji: 🚀
colorFrom: indigo
colorTo: red
pinned: true

Cardiovascular Disease Risk Predictor

This is a Streamlit-based web app to predict cardiovascular disease (CVD) risk using a trained FT-Transformer model.

🧠 Model

The model is trained on health data with features like age, blood pressure, cholesterol, and lifestyle factors. It uses a decision threshold of 0.38 to determine high risk.

🚀 How to Use

1. Clone this repository or upload to Hugging Face Spaces.
2. Ensure the folder structure is preserved.
3. The app will automatically launch using streamlit run app.py.

📁 Folder Structure

FTTransformerApp/
├── app.py
├── requirements.txt
└── FTTransformerModel/
    ├── model.ckpt
    ├── config.yml
    ├── encoder.pkl
    ├── threshold.json
    ├── callbacks.sav
    ├── custom_params.sav
    ├── datamodule.sav
    └── exp_logger.sav

✅ Prediction Threshold

This app uses a probability threshold of 0.38 to classify as high risk for CVD.