Improved multi-scale kernel principal component analysis and its application for fault detection

Progress of data-driven process monitoring for nonlinear and non-Gaussian industry process

“…Compared to the neural network, Kernelbased methods of statistic studying theory represented by SVM(support vector machine) has better generalization ability, and become another very popular and important techniques after artificial neural networks in the field of face recognition, machine learning and pattern recognition and get a wealth of research results. In recent years, kernel learning methods have been introduced to the field of industry process monitoring, and with the combination of traditional MSPC methods, other new nonlinear MSPC methods were proposed such as nuclear PCA (Kernel PCA, KPCA), kernel PLS (Kernel PLS, KPLS), nuclear ICA (Kernel ICA, KICA) [20][21][22]. Although Kernel-based methods have solved many difficulties exist in methods of the neural network, however, there is still a lot of questions, such as the choice of kernel function and computational efficiency in case of a large sample and so on.…”

“…[14][15][16][17][18][19][20][21][22][23]. In which, principal curves theory put forward by Hasti is actually nonlinear PCA method extended from a linear PCA, Camacho, Kampjarvi and Ghats also have put forward several nonlinear MSPC methods based on neural network for nonlinear process monitoring.…”

Multivariate data modeling using modified kernel partial least squares

“…To solve this issue, a nonlinear PLS method, called kernel partial least squares (KPLS), was proposed by Rosipal and Trejo (Rosipal and Trejo, 2002). The original datasets are nonlinearly transformed into a feature space of arbitrary dimensionality via nonlinear mapping, and then a linear model is created in the feature space (Zhang et al, 2012;Zhang and Hu, 2011). Because it's easy to understand and operate, KPLS has been widely used in many fields, such as pattern recognition (Qu et al, 2010), signal processing (Helander et al, 2012), fault diagnosis (Zhang et al, 2010b), and so on.…”

Fault detection for a class of industrial processes based on recursive multiple models

“…In the past decades, people focused on multivariate statistical process control (MSPC) [1] [2][3] [4]. Conventional MSPCs such as principal component analysis (PCA) and partial least squares yield satisfactory monitoring results for processes with a single operating mode without time varying under the assumption that the relationship between each variable is linear [5] [6].…”