L-SVM: A radius-margin-based SVM algorithm with LogDet regularization

作者：

Highlights：

• A negative LogDet regularization is introduced to the radius-based SVM algorithm.

• We also develop an efficient algorithm to solve our proposed L-SVM model.

• Experimental results show that our L-SVM model achieves better performance.

• We propose a fraud detection system and conduct a simulation experiment.

摘要

•A negative LogDet regularization is introduced to the radius-based SVM algorithm.•We also develop an efficient algorithm to solve our proposed L-SVM model.•Experimental results show that our L-SVM model achieves better performance.•We propose a fraud detection system and conduct a simulation experiment.

论文关键词：Support vector machine,Radius-margin ratio,Error bounds,LogDet regularization,Fraud detection system

论文评审过程：Received 13 June 2017, Revised 1 February 2018, Accepted 2 February 2018, Available online 10 February 2018, Version of Record 19 March 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.02.006