论文部分内容阅读
通过改进模糊聚类方法确定模糊模型的前件结构 ,然后对模糊推理关系矩阵进行QR分解 ,通过分析秩的亏损来确定聚类规则的有效性 ,然后采用基于矩阵UD分解最小二乘确定模糊模型的后件参数 ,实现模糊模型的结构和参数的优化 .该方法成功地应用于Box Jenkins煤气炉数据系统建模
The former structure of the fuzzy model is determined by improving the fuzzy clustering method, and then the QR-decomposition of the fuzzy inference matrix is performed. The validity of the clustering rules is determined by analyzing the rank loss. Then, the fuzzy model based on matrix UD decomposition least squares Of the back-piece parameters to achieve the fuzzy model structure and parameter optimization.This method has been successfully applied to Box Jenkins gas stove data system modeling