基于相似关系粗糙集中的否定规则及约简  

Negative Rules and Reduction in Rough Sets Based on Similarity Relation

作　　者：徐维艳[1] 魏敏[2] 张明[3] 

机构地区：[1]江苏科技大学数理学院,江苏镇江212003 [2]南京农业大学理学院,江苏南京210095 [3]江苏科技大学计算机学院,江苏镇江212003

出　　处：《微电子学与计算机》2012年第11期53-57,共5页Microelectronics & Computer

基　　金：国家自然科学基金项目(61100116)

摘　　要：为了从不完备决策信息系统中获取否定决策规则,首次提出了基于相似关系的不相似知识颗粒的概念,研究了如何通过不相似知识粒获取否定决策规则的方法,给出了知识约简的启发式算法,并进行了实例分析,为从不完备信息系统中获取知识提供了新的理论基础与操作手段.To derive the negative rules from the incomplete decision information system,the concept of dissimilarity knowledge granule based on similarity relation is firstly proposed.Moreover,the approach to acquire the certain and uncertain negative rules is presented,and heuristic algorithm of the attribute reduction based on the significant attributes is introduced.Finally,an illustrative example is analyzed.These results are meaningful both in the theory and applications for rules＇ acquisition in incomplete decision systems.

关 键 词：相似关系 知识获取 决策规则 约简 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]