首页 | 本学科首页   官方微博 | 高级检索  
     

基于INLR-PPLS的非线性多传感耦合信息建模新方法
引用本文:洪晓斌;刘桂雄. 基于INLR-PPLS的非线性多传感耦合信息建模新方法[J]. 华南理工大学学报(自然科学版), 2009, 37(8)
作者姓名:洪晓斌  刘桂雄
作者单位:华南理工大学机械与汽车工程学院;华南理工大学机械工程学院
摘    要:针对线性PLSR在多传感信息回归建模中存在不足,提出了一种基于INLR-PPLS的非线性多传感耦合信息建模新方法。首先用线性PLSR对传感信息进行预处理,达到降维的效果,然后基于INLR外模型非线性样本矩阵变换、基于PPLS内模型非线性映射、多传感非线性回归模型反求,实现对非线性多传感耦合信息的建模。应用于液态乙醇浓度测控系统的实验表明,该方法较线性PLSR预测准确度提高21%。

关 键 词:多传感信息  PLSR  INLR  PPLS  
收稿时间:2008-08-03
修稿时间:2008-09-12

Modelling Approach of Nonlinear Multi-sensor Coupling Information Based on INLR-PPLS
Xiao-bin HONGLIU Gui-Xiong. Modelling Approach of Nonlinear Multi-sensor Coupling Information Based on INLR-PPLS[J]. Journal of South China University of Technology(Natural Science Edition), 2009, 37(8)
Authors:Xiao-bin HONGLIU Gui-Xiong
Abstract:In allusion to deficiency of linear PLSR for multi-sensor information modelling, a novel modelling approach of nonlinear multi-sensor coupling information based on INLR-PPLS is put forward. The multi-sensor information is preprocessed first by means of linear PLSR to descend dimensions, then integrated INLR to construct outer model through nonlinear sample matrix transform and PPLS to form inter model by nonlinear mapping. Finally, the reverse regression model is accomplished. From the experiment of alcohol concentration measurement, the results indicate that the prediction accuracy of proposed approach is raised 21% at average, comparing to the linear PLSR in the similar industrial circumstances.
Keywords:multi-sensor information  PLSR  INLR  PPLS
点击此处可从《华南理工大学学报(自然科学版)》浏览原始摘要信息
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号