A relative decision entropy-based feature selection approach
作者：
Highlights：
• We proposed a novel heuristic feature selection algorithm in rough sets.
• We presented a new information entropy model – relative decision entropy.
• We proved that relative decision entropy is monotonic with respect to the partial order of partitions.
• We applied our feature selection algorithm to intrusion detection.
• The effectiveness of our algorithm was shown on KDD-99 data set and some other data sets.
摘要
Highlights•We proposed a novel heuristic feature selection algorithm in rough sets.•We presented a new information entropy model – relative decision entropy.•We proved that relative decision entropy is monotonic with respect to the partial order of partitions.•We applied our feature selection algorithm to intrusion detection.•The effectiveness of our algorithm was shown on KDD-99 data set and some other data sets.
论文关键词：Rough sets,Feature selection,Roughness,The degree of dependency,Relative decision entropy,Feature significance
论文评审过程：Received 31 August 2013, Revised 23 January 2015, Accepted 24 January 2015, Available online 4 February 2015.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.01.023
