集值信息系统中基于证据理论的规则提取  

Rule extraction in set-valued decision information system based on evidence theory

作　　者：田冉[1] 王国胤[1] 陶春梅[1] 

机构地区：[1]重庆邮电大学计算机科学技术学院,重庆400065

出　　处：《重庆邮电大学学报（自然科学版）》2011年第5期635-640,共6页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：国家自然科学基金(6057306860773113);重庆市杰出青年科学基金(2008BA2041);重庆市自然科学基金重点项目(2008BA2017)~~

摘　　要：为了提高在集值决策信息系统中最优广义决策规则的获取速度,借助证据理论这一处理不确定性问题的有力工具,将证据理论应用到集值信息系统中,定义了集值信息系统在相似关系和优势关系下的基本概率分配函数,提出一种基于基本概率分配函数的最优广义决策规则获取方法。最后用实例说明了这种方法的有效性,并且有效地降低了时间复杂度。In order to improve access speed of the optimal generalized decision rules from set-valued information system, this paper use evidence theory as an effective tool in dealing with uncertainty questions. This paper discusses the applica- tion of evidence theory in set-valued information systems, defines basic probability assignment functions on similarity rela- tion and dominance relation in set-valued information systems, proposes a new approach to extract optimal generalized deci- sion rules based on basic probability assignment. An example is given to prove the validity of this approach and it also shows the less time complexity of this approach than other methods.

关 键 词：证据理论 集值信息系统 粗糙集 决策规则提取 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]