模糊模型辨识中模糊聚类方法应用分析  

Analysis on the Applications of Fuzzy Clustering Method in Fuzzy Model Identification

作　　者：刘福才[1] 朴春俊[2] 裴润[3] 

机构地区：[1]哈尔滨工业大学控制工程系,黑龙江哈尔滨150001 [2]燕山大学电气工程学院自动化系,河北秦皇岛066004 [3]哈尔滨工业大学航天学院控制工程系,黑龙江哈尔滨150001

出　　处：《系统工程与电子技术》2002年第5期35-37,64,共4页Systems Engineering and Electronics

基　　金：黑龙江省自然科学基金资助课题;哈尔滨工业大学科学研究基金资助课题

摘　　要：模糊聚类算法已广泛应用于模式识别、数据聚类以及从数据中提取模糊规则的过程。介绍了基于模糊聚类的非线性系统模糊辨识方法 ,并通过著名的Box和Jenkins煤气炉数据仿真实例详细研究了模型性能指标与输入变量及模糊聚类数之间的关系 ,指出了应用模糊聚类方法的优势与不足。The fuzzy clustering algorithm has been extensively used for pattern recognition and data clustering. It has also been applied in the process of generating fuzzy rules from data. This paper introduces a nonlinear system fuzzy identification method based on fuzzy clustering, and studies the relation between the model is performance index and the number of input variables and fuzzy clustering centers in detail through the simulation of the famous Box-Jenkins gas furnace data. It also presents the advantages and disadvantages of using fuzzy clustering method. It is of great importance for the applications of fuzzy clustering method in fuzzy modeling.

关 键 词：模糊模型辨识 模糊聚类方法 模式识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]