Skip to content

Four Federated learning datasets for Human Activity Recognition

License

Notifications You must be signed in to change notification settings

xmouyang/FL-Datasets-for-HAR

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

49 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

FL-Datasets-for-HAR

This repo includes four new real-world human activity recognition (HAR) datasets collected under federated learning settings, which first appear at the MobiSys 2021 paper: ClusterFL: A Similarity-Aware Federated Learning System for Human Activity Recognition .

The first dataset is a large-scale dataset collected using an Android App in a crowdsourcing manner. The other three are collected in indoor environments.

Download

The four datasets are publicly available in the dropbox link . Each dataset is accompanied by a Python file "data_pre.py" for preprocessing and loading each node's data separately in federated learning. The data and processing file are compressed to a ".zip" file for each dataset. Please click the following links for more detail descriptions and downloading each dataset.

Citation

The datasets of this project are made available for non-commercial, academic research only. If you would like to use the datasets of this project, please cite the following papers:

@inproceedings{ouyang2021clusterfl,
  title={ClusterFL: a similarity-aware federated learning system for human activity recognition},
  author={Ouyang, Xiaomin and Xie, Zhiyuan and Zhou, Jiayu and Huang, Jianwei and Xing, Guoliang},
  booktitle={Proceedings of the 19th Annual International Conference on Mobile Systems, Applications, and Services},
  pages={54--66},
  year={2021}
}
@article{ouyang2022clusterfl,
  title={ClusterFL: A Clustering-based Federated Learning System for Human Activity Recognition},
  author={Ouyang, Xiaomin and Xie, Zhiyuan and Zhou, Jiayu and Xing, Guoliang and Huang, Jianwei},
  journal={ACM Transactions on Sensor Networks},
  volume={19},
  number={1},
  pages={1--32},
  year={2022},
  publisher={ACM New York, NY}
}

About

Four Federated learning datasets for Human Activity Recognition

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published