This is the code repository for the paper "Significance-aware Medication Recommendation with Medication Representation Learning".
python==3.9.18
torch==2.1.1
tqdm==4.66.1
dgl==1.1.2.cu118
scikit-learn==1.3.2
You can build the conda environment for our experiment using the following command:
conda env create -f environment.yml
All same as SafeDrug or COGNet except process.py
.
You can train or test the model using the following command:
python main.py
python main.py --test True
If this work is useful in your research, please cite our paper.
@inproceedings{li2024mr,
title={Significance-aware Medication Recommendation with Medication Representation Learning},
author={Yishuo Li, Zhufeng Shao, Weimin Chen, Shoujin Wang, Yuehan Du, Wenpeng Lu},
booktitle={Proceedings of the 27th International Conference on Computer Supported Cooperative Work in Design (CSCWD 2024)}
year={2024},
organization={IEEE}
}