Can high-order dependencies improve mutual information based feature selection?

作者：

Highlights：

• We shed light on the assumptions behind many popular mutual information based feature selection approaches.

• We propose a theoretically sound approach to extend current mutual information based feature selection methods to handle high-order dependencies.

• We provide concrete experimental evidence that high-order dependencies improve mutual information based feature selection.

摘要

Highlights•We shed light on the assumptions behind many popular mutual information based feature selection approaches.•We propose a theoretically sound approach to extend current mutual information based feature selection methods to handle high-order dependencies.•We provide concrete experimental evidence that high-order dependencies improve mutual information based feature selection.

论文关键词：Feature selection,Mutual information,High-order dependency

论文评审过程：Received 10 March 2015, Revised 7 September 2015, Accepted 10 November 2015, Available online 19 November 2015, Version of Record 8 February 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2015.11.007