Improving experimental studies about ensembles of classifiers for bankruptcy prediction and credit scoring

作者：

Highlights：

• We study previous works about datasets on bankruptcy prediction and credit scoring.

• We propose a procedure to improve the results of previous works.

• We show that Bagging scheme must be used on the correct base procedures.

• Our proposal is simpler and improves complex procedures.

摘要

•We study previous works about datasets on bankruptcy prediction and credit scoring.•We propose a procedure to improve the results of previous works.•We show that Bagging scheme must be used on the correct base procedures.•Our proposal is simpler and improves complex procedures.

论文关键词：Bankruptcy prediction,Credit scoring,Ensembles of classifiers,Decision trees,Imprecise Dirichlet model

论文评审过程：Available online 18 December 2013.

论文官网地址：https://doi.org/10.1016/j.eswa.2013.12.003