Title: Application of principal component analysis for mechanical coupling system modelling based on support vector machine

Authors: Jianwei Ma, Fuji Wang, Zhenyuan Jia, Wei Liu

Addresses: Key Laboratory for Precision and Non-traditional Machining Technology of the Ministry of Education, Dalian University of Technology, Liaoning Dalian 116024, China. ' Key Laboratory for Precision and Non-traditional Machining Technology of the Ministry of Education, Dalian University of Technology, Liaoning Dalian 116024, China. ' Key Laboratory for Precision and Non-traditional Machining Technology of the Ministry of Education, Dalian University of Technology, Liaoning Dalian 116024, China. ' Key Laboratory for Precision and Non-traditional Machining Technology of the Ministry of Education, Dalian University of Technology, Liaoning Dalian 116024, China

Abstract: This paper presents the results of a research into the application of principal component analysis (PCA) for the mechanical coupling system modelling based on support vector machine (SVM). Because of the impact of multiple geometric parameters, there are more input variables in the mechanical coupling system modelling process. The high-dimensional data poses an interesting challenge to machine learning, as the presence of high numbers of redundant or highly correlated variables can seriously degrade modelling accuracy. In this study, we use PCA as the preprocessor for mechanical coupling system modelling, so as to realise dimension reduction of the high-dimensional data and improve the predictive performance of machine learning method, and then SVM is used for mechanical coupling system modelling. Experiments are carried out on a typical mechanical coupling, hydraulic valve. The results show that the use of PCA method can improve the performance of machine learning method in the modelling of high-dimensional data.

Keywords: principal component analysis; PCA; support vector machines; SVM; hydraulic valves; characteristics prediction; mechanical coupling systems; modelling; machine learning.

DOI: 10.1504/IJMA.2011.040037

International Journal of Mechatronics and Automation, 2011 Vol.1 No.2, pp.71 - 78

Published online: 26 Mar 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article