UniTS-Sensory-Time-Series-Classification

Datasets and Pytorch code for UniTS: Short-Time Fourier Inspired Neural Networks for Sensory Time Series Classification, in The 19th ACM Conference on Embedded Networked Sensor Systems (SenSys 2021).

Prerequisite

PyTorch >= 1.8.0

numpy

scikit-learn

pytorch-complex

How to run

1. Run main.py for UniTS and the following baseline models (ResNet, MaCNN, RFNet-base, THAT, LaxCat)
2. Run complex_main.py for model STFNets.

Datasets

Processed:

https://drive.google.com/file/d/1aPb-iy6ic-bcg-azXVQ2_2uegn0oQc_j/view?usp=sharing

Raw:

Motion:

https://archive.ics.uci.edu/ml/datasets/opportunity+activity+recognition#:~:text=Data%20Set%20Information%3A-,The%20OPPORTUNITY%20Dataset%20for%20Human%20Activity%20Recognition%20from%20Wearable%2C%20Object,%2C%20feature%20extraction%2C%20etc).

Seizure:

https://physionet.org/content/chbmit/1.0.0/

WiFi:

https://github.com/ermongroup/Wifi_Activity_Recognition

KETI:

https://github.com/Shuheng-Li/Relational-Inference/tree/master/KETI_oneweek

Notes

Code will use cuda by default. You may tune model hyperparameter for better results.

Citations

Please cite the following paper if you use this repository in your research work:

@inproceedings{10.1145/3485730.3485942,
author = {Li, Shuheng and Chowdhury, Ranak Roy and Shang, Jingbo and Gupta, Rajesh K. and Hong, Dezhi},
title = {UniTS: Short-Time Fourier Inspired Neural Networks for Sensory Time Series Classification},
year = {2021},
isbn = {9781450390972},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3485730.3485942},
doi = {10.1145/3485730.3485942}
}

Contact Shuheng Li ✉️ for questions, comments and reporting bugs.