基于粗糙集理论的不完备信息处理方法研究  

Processing incomplete information methods based on rough set

作　　者：官礼和[1,2,3] 

机构地区：[1]西南交通大学信息科学与技术学院,成都610031 [2]重庆邮电大学计算机科学与技术研究所,重庆400065 [3]重庆交通大学光电与信息技术研究所,重庆400074

出　　处：《重庆邮电大学学报（自然科学版）》2009年第4期461-466,共6页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：国家自然科学基金(60573068);重庆市杰出青年科学基金(2008BA2041);重庆市教委科学技术研究项目(KJ090512);重庆交通大学青年科学基金(2007)

摘　　要：针对不完备信息处理问题,分析了不完备信息的语义解释,将粗糙集理论中处理不完备信息的技术策略归纳为数据补齐法和模型扩展法两类。在对已有典型粗糙集扩展模型的定义及其优缺点进行分析的基础上,分别对限制容差关系和特征关系的定义进行改进,得到新的扩充关系模型,讨论了它们的性质。对量化容差关系中对象间的相似度度量方法,给出了3种不同的度量方式及其相应的量化容差关系模型。Aiming at the problem of incomplete information processing, the semantics of incomplete information were analyzed, and the strategies for processing incomplete information in rough set theory were reduced to two kinds： data reparation and model extension. The definitions and characters of existing extended models were analyzed. On this basis, the limited tolerance relation and the characteristic relation were improved, and their characters were discussed. For the depiction methods of the objects similitude in valued tolerance relation, three different methods and their corresponding valued tolerance relations were proposed.

关 键 词：不完备信息 粗糙集 容差关系 特征关系 量化容差关系 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]