基于满意聚类的多模型建模方法

从系统输入输出数据出发, 首先在GK模糊聚类算法的基础上, 提出一种模糊满意聚类算法, 该算法能快速对系统进行用户满意的模糊划分 ;继而将其引入多模型建模过程中, 满意的系统划分数目即对应多模型个数, 然后针对不同的聚类建立起相应的子系统模型, 全局系统可视为各子模型的加权组合 ;最后通过几个典型实例验证了模糊满意聚类及基于此的多模型建模方法的有效性、准确性和快速性.

Multi-model modeling method based on satisfactory clustering

First of all, from input-output data set, a satisfactory clustering algorithm based on GK fuzzy clustering was presented. Using this algorithm, a system could be quickly divided into multiple optimal fuzzy parts. Then the algorithm was used in multi-model modeling process. Satisfactory cluster number corresponded to the optimal number of sub-systems. For multiple clusters, multiple models could then be built, and the global system was described as their certain combination. Finally, examples are given to prove the effectiveness of the method.