Reject inference in credit scoring using Semi-supervised Support Vector Machines
作者：
Highlights：
• Semi-supervised Support Vector Machines for reject inference are proposed.
• The method uses information of both the accepted and rejected applicants.
• The method deals with labelled and unlabelled classes of the outcome.
• The model is tested on real consumer loans with a low acceptance rate.
• Predictive accuracy is improved by the new model compared to traditional methods.
摘要
•Semi-supervised Support Vector Machines for reject inference are proposed.•The method uses information of both the accepted and rejected applicants.•The method deals with labelled and unlabelled classes of the outcome.•The model is tested on real consumer loans with a low acceptance rate.•Predictive accuracy is improved by the new model compared to traditional methods.
论文关键词：Reject inference,Credit scoring,Semi-supervised Support Vector Machines,Online lending,Predictive accuracy
论文评审过程：Received 2 October 2016, Revised 21 December 2016, Accepted 11 January 2017, Available online 12 January 2017, Version of Record 19 January 2017.
论文官网地址：https://doi.org/10.1016/j.eswa.2017.01.011
