Parametric methods for comparing the performance of two classification algorithms evaluated by k-fold cross validation on multiple data sets

作者：

Highlights：

• Parametric methods for performance comparison of two classification algorithms are proposed.

• Both independent and matched samples are considered.

• The experimental results on twelve data sets demonstrate the effectiveness of our parametric methods.

摘要

•Parametric methods for performance comparison of two classification algorithms are proposed.•Both independent and matched samples are considered.•The experimental results on twelve data sets demonstrate the effectiveness of our parametric methods.

论文关键词：Classification,k-fold cross validation,Parametric method,Sampling distribution

论文评审过程：Received 22 June 2016, Revised 25 November 2016, Accepted 15 December 2016, Available online 16 December 2016, Version of Record 25 December 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2016.12.018