摘要
在数据挖掘、模糊专家系统和多Agent协同决策过程中,要经常面对信息聚集技术和对多个模糊数据来源进行聚集运算,一般用得最多的是合取、析取及加权平均等算子,但是不同的领域有着不同的需求,本文着重对加权聚集算子进行研究。首先,提出了加权平均关系与析取关系结合后的析取-加权平均算子,该算子解决了加权平均算子不能区分析取与合取的关系。然后,提出了一种最大加权平均算子和最小加权平均算子,该算子将最大最小值算子与加权平均算子进行了泛化,解决了同时考虑信息局部性特征与信息整体性特征的问题。理论分析表明,本文提出的加权模糊聚集算子对于模糊信息源的聚集运算起到了很好的补充和完善的功能。
Information fusion techniques and fuzzy aggregation operators are commonly applied into several fields of data mining, fuzzy expert system and multi-agent cooperated decision. Because different fields imply different requirements,a large number of aggregation operators exist today. The mostly used is Maximum operators, Minimum operations, and weighted mean operators. In particular, this article pays more attention on weighted fuzzy operators. First, we put forward the disjunction-weighted average operators, which solve the problem that weighted average operator cannot distinguish the relationship between disjunction and conjunction. Secondly, this article generalizes the maximum/minimum and weighted average operators, puts forward maximum weighted average operators and minimum weighted average operators. The fuzzy information sources local character and whole character are both considered in these operators. The theory results show these fuzzy aggregation operators are good complements for fuzzy information fusion.
出处
《计算机科学》
CSCD
北大核心
2007年第2期196-197,203,共3页
Computer Science
基金
国家"八六三"高技术研究发展计划基金项目(2004AA112020
2005AA112030)
武器装备预研基金项目(51415010304KG0175)
关键词
模糊聚集算子
析取-加权平均算子
加权算子
最大加权平均算子
最小加权平均算子
Fuzzy aggregation operators, Disjunction-weighted average operators, Weighted operators, Maximum weighted average operators, Minimum average operators
