A feature selection algorithm based on approximate markov blanket and dynamic mutual information

  • Authors:

  • Xiaodan Wang;Xu Yao;Yuxi Zhang;Lei Lei

  • Affiliations:

  • Department of Computer Engineering, Air Force Engineering University, China;Department of Computer Engineering, Air Force Engineering University, China;Department of Computer Engineering, Air Force Engineering University, China;Department of Computer Engineering, Air Force Engineering University, China

  • Venue:
  • IScIDE'11 Proceedings of the Second Sino-foreign-interchange conference on Intelligent Science and Intelligent Data Engineering
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Based on the research on feature relevance, features can be divided into four categories: strong relevance, weak relevance, irrelevance and redundancy. Feature selection is a process of removing irrelevance and redundancy features in nature. A feature selection algorithm is given, which uses dynamic mutual information as evaluation criteria and eliminates irrelevance and redundancy features by approximate Markov Blanket. Experimental results on UCI data sets with support vector machine as the classifier indicate the feasibility and validity of the algorithm proposed in this paper.