上海交通大学学报(自然版) ›› 2017, Vol. 51 ›› Issue (6): 722-726.

• 兵器工业 • 上一篇    下一篇

 一种基于自适应模糊高斯核聚类的软测量建模方法

 夏源,杨慧中   

  1.  江南大学 轻工过程先进控制教育部重点实验室, 江苏 无锡 214122
  • 出版日期:2017-06-30 发布日期:2017-06-30
  • 基金资助:
     

 A Soft Sensor Modeling Method Based on
 SelfAdaptive Fuzzy Gauss Kernel Clustering

 XIA Yuan,YANG Huizhong   

  1.  Key Laboratory of Advanced Control in Light Industry Process of Ministry of Education,
    Jiangnan University, Wuxi 214122, Jiangsu, China
  • Online:2017-06-30 Published:2017-06-30
  • Supported by:
     

摘要:  单一模型一般难以表达复杂的生产过程特性,在软测量应用中往往容易使模型的估计精度低、泛化性能差.提出一种基于自适应模糊高斯核聚类的概率加权多模型融合方法,利用高维空间内样本的分散性来确定聚类中心,能取得最佳聚类效果.根据贝叶斯后验定律进行多模型融合,使总模型输出更具合理性.该方法不仅克服了单模型预测的局限性,同时对传统多模型融合方法做了一些改进,提高了过程估计的精度.

关键词:  , 自适应, 模糊高斯核聚类, 概率加权, 多模型

Abstract:  It is difficult for a single model to express the complicated production process, and it often results in low accuracy of prediction and poor performance of generalization. This paper presents a multimodel fusion method based on probability weight and selfadaptive fuzzy Gauss kernel clustering. The method determines cluster centers according to dispersion of the samples in a high dimensional space. The weight of every submodel is given by Bayesian posterior method. The method can overcome the limitation of singlemodel forecast and improve traditional multimodel fusion methods for obtaining higher prediction accuracy.

Key words:  adaption, fuzzy Gauss kernel clustering, probabilityweighted, multimodel

中图分类号: