A new hybrid ensemble credit scoring model based on classifiers consensus system approach
作者：
Highlights：
• A new hybrid ensemble model for credit scoring problem is proposed.
• An improved data filtering technique is developed based on GNG method.
• GNG with MARS combined proved to be better than applying them individually.
• Our model is validated on four performance measures over seven credit datasets.
• Classifiers decisions after consensus effectively improved prediction performance.
摘要
•A new hybrid ensemble model for credit scoring problem is proposed.•An improved data filtering technique is developed based on GNG method.•GNG with MARS combined proved to be better than applying them individually.•Our model is validated on four performance measures over seven credit datasets.•Classifiers decisions after consensus effectively improved prediction performance.
论文关键词：Credit scoring,Consensus approach,Classifier ensembles,Hybrid models,Data filtering,Feature selection
论文评审过程：Received 28 March 2016, Revised 10 July 2016, Accepted 11 July 2016, Available online 14 July 2016, Version of Record 27 July 2016.
论文官网地址：https://doi.org/10.1016/j.eswa.2016.07.017
