摘要
首先在常规的Bezdek模糊划分模型的基础上,引入了考虑权重的模糊划分模型;并根据数理统计中的方差分析理论和信息论中的信息熵理论,构建了模糊F统计量、模糊划分熵,F统计量用来确定最佳划分数,模糊划分熵用来检验模糊划分的有效性.最后通过实际算例对这两个函数的判决功能和F统计量的鲁棒性进行了研究.
On the base of normal Bezdek fuzzy division model, the new fuzzy division model considering weights was introduced. According to the theory of difference analysis in statistics and entropy theory in information theory, two functions-F statistics and fuzzy division entropy were constructer, the F statistics was mainly used to determine the optimal division number and the fuzzy division entropy was used to verify the validity of division. At last, a practical example was presented to verify the effective of the two functions and robust property of F statistics.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2004年第2期106-110,共5页
Systems Engineering-Theory & Practice
基金
国家重点基础发展规划研究项目(G1999043606)
国家社会科学基金(03CJL006)
关键词
模糊划分
权重
方差分析
信息熵
F统计量
模糊划分熵
有效性
鲁棒性
fuzzy division
weights
difference analysis
information entropy
F statistics
fuzzy division entropy
validity
robust property
