Accelerating wrapper-based feature selection with K-nearest-neighbor
作者：
Highlights：
• We propose to accelerate wrapper-based feature selection with a KNN classifier.
• We construct a classifier distance matrix to evaluate the quality of a feature.
• The proposed approach can apply to three types of wrapper-based feature selectors.
• Theoretical time complexity analysis proves the efficiency of the proposed approach.
• Experimental results demonstrate its effectiveness and efficiency.
摘要
•We propose to accelerate wrapper-based feature selection with a KNN classifier.•We construct a classifier distance matrix to evaluate the quality of a feature.•The proposed approach can apply to three types of wrapper-based feature selectors.•Theoretical time complexity analysis proves the efficiency of the proposed approach.•Experimental results demonstrate its effectiveness and efficiency.
论文关键词：Gene selection,Microarray data,Wrapper,Filter,k-nearest-neighbor
论文评审过程：Received 12 November 2014, Revised 11 March 2015, Accepted 13 March 2015, Available online 21 March 2015.
论文官网地址：https://doi.org/10.1016/j.knosys.2015.03.009
