code for paper Skyline of Counterfactual Explanations (CIKM 2021)
pip install numpy, pandas, pytorch, sklearn
- The cf/ folder contains the a series of CF methods, includign skylineCF in our paper.
- data/ folder contains the datasets used.
- model folder consists of the interface for loading the pretrained model.
- util folder includes the interface for loading data, skyline operator used in our paper.
- weights folder covers the weights of pretrained model on three datasets.
- test folder comprise the how to generate the counterfactual explaantions for our method and baselines.
- mlp_* files show how to train the models.
@inproceedings{wang2021skyline,
title={The Skyline of Counterfactual Explanations for Machine Learning Decision Models},
author={Wang, Yongjie and Ding, Qinxu and Wang, Ke and Liu, Yue and Wu, Xingyu and Wang, Jinglong and Liu, Yong and Miao, Chunyan},
booktitle={Proceedings of the 30th ACM International Conference on Information \& Knowledge Management},
pages={2030--2039},
year={2021}
}