摘要
本文探讨了基于属性重要性、基于信息熵、基于遗传算法和基于聚类的离散化算法,通过分析总结了各算法的优点及不足,并提出有待解决的问题。
Four kinds of discrctization algorithms are discussed.They are significance of attributes, information entropy, genetic algorithm and clusteringbased algorithms. The merits, shortages and some pending problems of these algorithms are addressed by analysis.
基金
吉林省教育厅科技计划基金资助项目(吉教科合字[2007]第172号)
黑龙江省教育厅科学技术研究项目(11531390)
关键词
离散化
属性重要性
信息熵
遗传算法
聚类
discretization
significance of attributes
information entropy
genetic algorithm
clustering