Attribute Reduction of Rough Set Based on Particle Swarm Optimization with Immunity
Weihua Lin, Yonggang Wu, Dianhui Mao, Yan Yu · 2008
Since the amount of spatial data grows rapidly during recent years, high dimensionality of the domain attributes presents a further obstacle for a number of rule-induction algorithms that would have the potential for automating knowledge acquisition. This paper attempts to tackle the problem by attribute reduction. Firstly, the problem of attribute reduction can be converted into a 0-1 combinatorial optimization problem after the theory of rough set are introduced. Secondly, a binary particle swam optimization algorithm with immunity (BPSOI) is proposed to deal with the problem. Thirdly, an experiment of attribute reduction by constructing data is presented, which indicates that the new algorithm is a very effective method to solve the problem of attribute reduction. Therefore, the new algorithm can improve the efficiency of attribute reduction by a long way, which is employed to remove the redundant and information-poor attributes in the field of knowledge discovered or data mining.