摘要
针对许多复杂系统的输入变量之间存在的相互关联,提出了一种基于聚类与模糊关联规则的神经模糊建模方法.这种方法采用基于聚类的模糊关联规则挖掘算法来进行输入变量的选择,之后,再采用基于减法聚类的神经模糊建模方法建模.最后,还将这种建模方法应用于实际建模问题,结果表明这种方法在保证模型精度符合建模要求的情况下,减少了模型输入个数,降低了建模的复杂程度.
The paper proposes a new neuro-fuzzy modeling method based on subtractive clustering and mining fuzzy association rules, in the light of the co-relations among input variables of many complicated systems, The method applies the algorithm of mining fuzzy association rules to select the input variables, and then constructs the model with the method of neuro-fuzzy modeling method based on subtractive clustering. The method has been applied to practice and the results show that, under the condition that the model accura- cy meets the modeling requirement, it has reduced model inputs and lowered the complexity of the modeling.
出处
《集美大学学报（自然科学版）》
CAS
2007年第1期59-63,共5页
Journal of Jimei University：Natural Science
基金
福建省自然科学基金资助项目(A0410005)
