An overview and a benchmark of active learning for outlier detection with one-class classifiers

作者：

Highlights：

• Categorization of assumptions and objectives of one-class active learning.

• Novel progress curve summaries to facilitate reliable evaluation of active learning.

• Large benchmark with 84,000 learning scenarios, classifiers, and query strategies.

• Derivation of guidelines to select suitable one-class active learning methods.

摘要

•Categorization of assumptions and objectives of one-class active learning.•Novel progress curve summaries to facilitate reliable evaluation of active learning.•Large benchmark with 84,000 learning scenarios, classifiers, and query strategies.•Derivation of guidelines to select suitable one-class active learning methods.

论文关键词：Active learning,One-class classification,Outlier detection,Anomaly detection

论文评审过程：Received 10 January 2020, Revised 23 November 2020, Accepted 23 November 2020, Available online 26 November 2020, Version of Record 14 December 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.114372