Abstract:
A method of fuzzy identification based on genetic soft fuzzy clustering and Kalman filtering method is proposed.The algorithm has three steps:(1) identify the optimal the number of clusters of Fuzzy c-mean (FCM) based on genetic algorithm,which can identify the conditional part of the rules and calculate the membership degree of sample data belonging to each cluster;(2) initialize parameter of fuzzy model by LS method;(3) the parameters of consequent part of the rules are tuned by Kalman filtering method.At last,the paper uses the algorithm to extracting the fuzzy rules and the result is shown.
摘要: 提出了一种新的基于遗传模糊软分类和卡尔曼滤波方法的模糊辨识算法。该算法由3个步骤组成:(1)基于遗传算法确定模糊C均值(FCM)中的最佳分类数,从而确定模糊规则的前件和样本在各类中的隶属度;(2)采用最小二乘法(LS)来确定模糊规则后件的初始参数;(3)用卡尔曼滤波方法调整后件参数。最后,运用该算法对我国全要素生产力进行了模糊规则的提取。
关键词:
卡尔曼滤波法,
模糊模型,
模糊规则,
遗传算法
CLC Number:
GUO Haixiang,ZHU Kejun,LI Jinling. Fuzzy Rules Extraction Based on Soft Clustering and Kalman Filter Method[J]. , 2005, 31(18): 201-203.
郭海湘,诸克军,黎金玲. 基于软分类和卡尔曼滤波方法的模糊规则提取[J]. , 2005, 31(18): 201-203.