基于矩阵QR分解的模糊模型结构分析与简化

Analysis and Simplification for Fuzzy Model Structure Based on Matrix QR Decomposition

研究用矩阵ＱＲ分解分析一类模糊模型的结构，用模糊统计信息判据指导规则的精简，用Ｒ矩阵的主对角线元素分布与模糊规则的关系确定重要规则、冗余或不重要规则，实现模糊模型结构优化。该方法已成功地应用于Ｍａｃｋｅｙ－

Analyzing model structure for a class of fuzzy models with matrix QR decomposition and guiding fuzzy rule simplification using fuzzy statistical information criteria are presented. According to the relation between the fuzzy rules and the value distribution of the primary diagonal elements in the matrix R, the more important rules and unimportant or redundant rules in the model can be determined respectively, and then an optimal fuzzy model can be constructed. The method has been applied successfully in Mackey-Glass chaotic system modeling.