BINet: Multi-perspective business process anomaly classification

作者：

Highlights：

• BINet neural network architecture.

• Automatic threshold heuristic.

• Generation algorithm for synthetic event logs.

• Comprehensive evaluation of state-of-the-art methods.

• Method to classify anomalies based on outputs of BINet.

摘要

•BINet neural network architecture.•Automatic threshold heuristic.•Generation algorithm for synthetic event logs.•Comprehensive evaluation of state-of-the-art methods.•Method to classify anomalies based on outputs of BINet.

论文关键词：Business process management,Anomaly detection,Artificial process intelligence,Deep learning,Recurrent neural networks

论文评审过程：Received 3 February 2019, Revised 29 June 2019, Accepted 21 October 2019, Available online 26 October 2019, Version of Record 27 September 2021.

论文官网地址：https://doi.org/10.1016/j.is.2019.101458