This is a candidate developed for the hackathon by team ATR21 during the semi-final round. This is a web-app based on Machine Learning models trained on cardio vascular and diabetic data, which can be used to predict the respective conditions based on user-provided inputs.
- Datasets
- Jupyter Notebooks for training models (in progress)
- Documentation
- Web-app
- The app and the ML models are developed using Anaconda Data Science platform 4.11.0
Requirements/dependencies are listed in
reqirements.yaml. - No configuration is needed unless you want to update the models with new data or further tune the model hyperparameters
- To tune the models, use the Jupyter Lab notebooks.
- Recreate the original development environment using:
conda env create -n <new env name> -f requirements.yaml - Clone or unzip this repository in your folder
- Run
healthpred.py. This will start a Flask web server on port 8080 - Access this app via your browser at
localhost:8080
- The webapp starting page has usage instructions