摘要
指出粗集理论的主要思想是在保持分类能力不变的情况下,利用等价类,通过属性约简和决策规则约简,达到挖掘知识并简化知识的目的.但约简问题是一个NP问题,只能通过启发式算法实现.针对这一问题,提出了属性约简和决策规则约简的启发式算法,构成了一个基于粗集理论的挖掘集成算法.最后通过实例表明,该集成算法能够以较高的效率发现良好的分类规则.
The essence of rough set theory rests with that it uses the equivalence relation class,attribute reduction and decision rule reduction to attain the object of knowledge mining and reduction of knowledge with unaltered power of classification.But the reduction is an NP problem,it can be solved only by method of elicitation.Aimed at this point,two methods of elicitation for attribute reduction and decision rule reduction were proposed,so that a mining integrate algorithm based on rough set theory was formed.Finally,an example showed that this integrate algorithm can be used to find better classification rules with a higher efficiency.
出处
《兰州理工大学学报》
CAS
北大核心
2005年第2期88-91,共4页
Journal of Lanzhou University of Technology
基金
甘肃省自然科学基金(ZS003 B35 026 C)
光电技术与智能控制教育部重点实验室开放基金(K04103)
关键词
粗集
决策系统
分辨矩阵
属性约简
数据挖掘
rough set
decision system
discrimination matrix
attribute reduction
data mine