TACCO

[KDD' 24] This repo is an official implementation of TACCO: Task-guided Co-clustering of Clinical Concepts and Patient Visits for Disease Subtyping based on EHR Data in PyTorch.

Model Pipeline

Prerequisites

- PyTorch 2.1.2
- scikit-learn 1.3.2
- torch-cluster 1.6.3
- torch_geometric 2.4.0
- torch-scatter 2.1.2
- torch-sparse 0.6.18
- pandas 2.1.4
- transformers 4.36.2

Data Preparation

We offer two randomly generated datasets in /data for training demonstration. Thus, their experimental results should NOT reflect the performance we report in the paper. For the authentic data we used, please request access to MIMIC-III.

Train TACCO

The script demo.sh is a ready-to-run shell script for a complete TACCO training process. We also provide the parameters we used in our paper, please check out KDD24.sh. The training results will be automatically saved in /logs.

Acknowledgments

This work partly uses the code from CACHE.

Cite

If you find our work useful, please cite our paper.

@article{zhang2024tacco,
  title={TACCO: Task-guided Co-clustering of Clinical Concepts and Patient Visits for Disease Subtyping based on EHR Data},
  author={Zhang, Ziyang and Cui, Hejie and Xu, Ran and Xie, Yuzhang and Ho, Joyce C and Yang, Carl},
  journal={arXiv preprint arXiv:2406.10061},
  year={2024}
}