This project is focused on using swarm learning and big data to build a model that can detect heart disease. The dataset used in this project is collected from Kaggle and is available at the following links:
- https://www.kaggle.com/datasets/vlbthambawita/deepfake-ecg
- https://www.kaggle.com/code/bjoernjostein/deepfake-ecg-generator
Swarm learning is a novel approach to machine learning that utilizes decentralized computing to enable data sharing and collaborative learning while maintaining privacy and data security. This approach is particularly well-suited for large-scale datasets where traditional machine learning methods may not be feasible.
The project's tech stack includes:
- Model CNN, SVM (Python)
- Backend: Flask
- Frontend: NuxtJS (VueJS)
By leveraging the power of swarm learning and big data, we aim to create a model that can accurately detect heart disease and improve patient outcomes.
Frontend - NuxtJS
- Install docker, follow the instruction in the directory /ecg-portal/docker-instruction.md
- Or running local, required: node v16.15.0 +
cd ecg-portal
yarn
yarn serve
- Access http://localhost:3000
Required: Install all packages which use by models
- Install python3
- Install joblib version 1.2.0
- tensorflow==2.12.0
- numpy
- sklearn
- pandas
cd backend
python3 main.py
Backend will run on http://localhost:5000