An outlier detection (anomaly detection) algorithm for categorical data.
This is source code of the method proposed in "Embedding-Based Complex Feature Value Coupling Learning for Detecting Outliers in Non-IID Categorical Data".
A sample dataset is in "data" folder. The expected dataset should be in csv format, and the attribute/feature is supposed to be categorical/nominal. main.py is used to perform and evalute the outlier detection process. You may also want to find sample usage of our method in main.py.
Please cite our paper if you find it is useful:
@inproceedings{xu2019embedding, title={Embedding-Based Complex Feature Value Coupling Learning for Detecting Outliers in Non-IID Categorical Data}, author={Xu, Hongzuo and Wang, Yongjun and Wu, Zhiyue and Wang, Yijie}, booktitle={Proceedings of the AAAI Conference on Artificial Intelligence}, volume={33}, pages={5541--5548}, year={2019} }