/Fed-TDA

The implementation of our paper Fed-TDA

Primary LanguagePythonMIT LicenseMIT

Fed-TDA

The implementation of our paper Fed-TDA: Federated Tabular Data Augmentation on Non-IID Data

Other baseline methods are as follow:

[1] FedMix: Approximation of Mixup under Mean Augmented Federated Learning| paper| code

[2]FEDERATED OPTIMIZATION IN HETEROGENEOUS NETWORKS |paper|code

[3]Fed-TGAN: Federated learning framework for synthesizing tabular data|paper|code

[4]Generative models for effective ML on private, decentralized datasets|paper|code

Usage Example

Run this repo:

  1. generate synthetic data on Clinical dataset:
python clinical_TDA_syn.py
  1. run script "clinical_eval.ipynb" to evaluate the performance of data augmentation

Citing Fed-TDA

@article{duan2022fed,
  title={Fed-TDA: Federated Tabular Data Augmentation on Non-IID Data},
  author={Duan, Shaoming and Liu, Chuanyi and Han, Peiyi and He, Tianyu and Xu, Yifeng and Deng, Qiyuan},
  journal={arXiv preprint arXiv:2211.13116},
  year={2022}
}