Skip to content

The official implementation of BIBM 24' Paper "Temporal Gaussian Copula For Clinical Multivariate Time Series Data Interpolation"

Notifications You must be signed in to change notification settings

MVL-Lab/TGC-MTS

Repository files navigation

Temporal Gaussian Copula For Clinical Multivariate Time Series Data Interpolation (BIBM 2024)

PRs Welcome Stars Visits Badge GitHub forks

Introduction

Multivariate time series (MTS) is particularly valuable in clinical diagnosis and disease surveillance. The imputation of the MTS is particularly applicable and challenging since the MTS typically contains irregular patterns of missing values, due to various factors such as instrument failures, interference from irrelevant data, and privacy regulations. Although existing statistical methods and deep learning methods have shown promising results in time series imputation, they failed to fully exploit the stronger temporal relationships among clinical variables when applied to the MTS. Moreover, they struggle to handle the MTS data variations in sampling density, a very common issue in MTS. To address these challenges, in this paper, we propose a Temporal Gaussian Copula Model (TGC) for three-order MTS imputation. The key idea is to leverage the Gaussian Copula to explore the cross-variable and temporal relationships based on the latent Gaussian representation. Subsequently, we employ an Expectation-Maximization (EM) algorithm to enhance the robustness of handling data with varying miss rates. We finally theoretically prove our proposed TGC is equivalent to the three-order Gaussian Copula. Comprehensive experiments were conducted on three real-world MTS datasets. The results demonstrate that our TGC substantially outperforms the state-of-the-art imputation methods. Additionally, the TGC model exhibits stronger robustness to the varying missing ratios in the test dataset.

method

Setup

pip install -r requirements.txt

Testing on PhysioNet2012

python TGC_PhysioNet2012.py

Acknowledgement

We would like to thank gcimpute and pypots for the codebase.

We have modified some of the code in the gcimpute package, so we put it in file list.

Citation

If you find this repo useful, please cite our paper.

@inproceedings{su2024temporal,
  title={Temporal Gaussian Copula For Clinical Multivariate Time Series Data Imputation},
  author={Su, Ye and Qiao, Hezhe and Wu, Di and Chen, Yuwen and Chen, Lin},
  booktitle={2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)},
  pages={3717--3721},
  year={2024},
  organization={IEEE}
}

About

The official implementation of BIBM 24' Paper "Temporal Gaussian Copula For Clinical Multivariate Time Series Data Interpolation"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •  

Languages