Python OAM: presentation and use of an outlying aspect mining library

• Rodrigo F. Silva Federal Technological University of Paraná (UTFPR)
• Luiz Gomes-Jr Federal Technological University of Paraná (UTFPR)

DOI: https://doi.org/10.5753/sbbd_estendido.2022.21846

Abstract

Outlier detection is used to identify system failures and fraud, among other applications. Detection algorithms are, however, limited in terms of providing information about the reason for the anomaly. Outlier Aspect Mining (OAM) assesses which aspects of the anomalous observation separate it from others. This article describes the implementation and use of a Python library that allows the user to apply OAM algorithms and analyze the results. We demonstrate the application of the library in a use case related to the COVID-19 pandemic.

Keywords: Anomaly Explainability, Outlier Detection, OAM, Python Library, Outlying Aspect Mining, Outiler, Python, Ipath

References

Chandola, V., Banerjee, A., and Kumar, V. (2009). Anomaly detection: A survey. ACM Computing Surveys, 41(3):15:1-15:58.

Cheng, Z., Zou, C., and Dong, J. (2019). Outlier detection using isolation forest and local outlier factor. In Proceedings of the Conference on Research in Adaptive and Convergent Systems, RACS ’19, pages 161-168, New York, NY, USA. Association for Computing Machinery.

Samariya, D., Aryal, S., and Ting, K. M. (2020a). A new effective and efficient measure for outlying aspect mining. arXiv: 2004.13550.

Samariya, D., Ma, J., and Aryal, S. (2020b). A Comprehensive Survey on Outlying Aspect Mining Methods. arXiv: 2005.02637.

the Vinh, N., Chan, J., Romano, S., Bailey, J., Leckie, C., Ramamohanarao, K., and Pei, J. (2016). Discovering outlying aspects in large datasets. Data Mining and Knowledge Discovery, 30.