基于粗集的多知识库决策融合  

Decision fusion of redundant knowledge based on rough set

作　　者：王加阳[1] 王国仁[2] 

机构地区：[1]中南大学信息科学与工程学院,长沙410083 [2]东北大学信息科学与工程学院,沈阳110004

出　　处：《控制与决策》2007年第6期657-662,共6页Control and Decision

基　　金：国家十五攻关计划项目(2002BA218C);湖南省自然科学基金项目(06JJ20075)

摘　　要：粗糙集理论研究的重要内容是分类与约简,其目的在于获取优良的规则知识,实现准确的决策.通过提出规则集合的决策度量思想,从整体上体现了对一个规则集合的衡量,为基于多知识库的决策奠定了基础.基于模型集成的基本思想,将规则知识库作为一个决策模型,根据规则集度量选择模型,通过模型集成实现决策融合.The important research contents of rough set are classification and reduction, the purpose of which is to get rule knowledge and realize accuracy decision. The metrics for rule set is presented and the properties of a rule set are showed in general, which play an important role for decision of redundant knowledge. Based on the basic theory of model integration rough set, a formal model for multi-repositories is provided. One repository is regarded as a decision model. By building a model integration procedure decision fusion is realized.

关 键 词：粗糙集 多知识库 决策融合 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]