基于混合不可分辨关系的粗糙集模型与知识获取  

Rough Set Model and Knowledge Acquisition Based on Hybrid Indiscernibility Relation

作　　者：徐苏平[1] 杨习贝[1] 范霁月[1] 陈才[1] 颜成[1] 

机构地区：[1]江苏科技大学计算机科学与工程学院,江苏镇江212003

出　　处：《淮海工学院学报（自然科学版）》2012年第4期28-32,共5页Journal of Huaihai Institute of Technology:Natural Sciences Edition

基　　金：国家自然科学基金资助项目(61100116);江苏省自然科学基金项目(BK2011492);江苏省高校自然科学基金项目(11KJB520004);中国博士后科学基金项目(20100481149);江苏省博士后科学基金项目(1101137C)

摘　　要：为了从决策系统中获取既有"并且"又有"或者"的决策规则,在强不可分辨关系和弱不可分辨关系的基础上,提出了混合不可分辨关系的概念。分别采用混合不可分辨关系和基于混合不可分辨关系的极大相容块构建了两种不同的粗糙集模型,并进行了对比分析。最后根据混合不可分辨关系极大相容块技术,讨论了如何从决策系统中获取决策规则并进行了可信度度量分析。To derive decision rules with both ＂AND＂ and ＂OR＂ explanations from the decision system, the concept of the hybrid indiscernibility relation was proposed through the fusion of strong and weak indiscernibility relations. Two different rough sets were constructed by emplo- ying hybrid indiscernibility relation and indiscernibility relation based on maximal consistent blocks. The relationships between such two different rough sets were then explored. Finally, based on the maximal consistent blocks technique, the generation of decision rules from decision system was addressed, and the certainty factor of the derived decision rules was considered.

关 键 词：强不可分辨关系 弱不可分辨关系 混合不可分辨关系 粗糙集 决策规则 可信度 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]