A hybrid method with dynamic weighted entropy for handling the problem of class imbalance with overlap in credit card fraud detection

作者：

Highlights：

• A Hybrid Method for the problem of data imbalance with overlap in fraud detection.

• Unsupervised anomaly detection model applied for obtaining overlapping subset.

• A novel Dynamic Weighted Entropy for guiding the sub-model selection.

• The proposed method outperforms others on real datasets.

摘要

•A Hybrid Method for the problem of data imbalance with overlap in fraud detection.•Unsupervised anomaly detection model applied for obtaining overlapping subset.•A novel Dynamic Weighted Entropy for guiding the sub-model selection.•The proposed method outperforms others on real datasets.

论文关键词：Class imbalance,Data overlap,Dynamic weighted entropy,Fraud detection

论文评审过程：Received 25 July 2020, Revised 4 January 2021, Accepted 16 February 2021, Available online 25 February 2021, Version of Record 13 March 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.114750