Abstract:Kernel principal component analysis (KPCA)-based process monitoring methods have recently shown to be very effective for monitoring nonlinear processes. However, their performances largely depend on the kernel function and currently there is no general rule for kernel selection. Existing methods simply choose the kernel function empirically or experimentally from a given set of candidates. This paper proposes a kernel function learning method for KPCA to learn a kernel function tailored to specific data and explores its potential for KPCA-based process monitoring. Motivated by the manifold learning method maximum variance unfolding (MVU), we obtain the kernel function by optimizing over a family of data-dependent kernels such that the nonlinear structure in input data is unfolded in the kernel feature space and gets more likely to be linear there. Using the optimized kernel, the nonlinear principal components of KPCA which are linear principal components in the kernel feature space can effectively capture the variation in data, and thus the data under normal operating conditions can be more precisely modeled by KPCA for process monitoring. Simulation results on an simple nonlinear system and the benchmark Tennessee Eastman (TE) demonstrate that the optimized kernel functions lead to significant improvement in the performance over the popular Gaussian kernels when used in the KPCA-based process monitoring.

Phase Space Reconstruction of Nonlinear Time Series Based on Kernel Method

Adaptive Kpca Modeling of Nonlinear Systems

Improved <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M1"><mml:mrow><mml:mi>k</mml:mi></mml:mrow></mml:math>-<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M2"><mml:mrow><mml:mi>t</mml:mi></mml:mrow></mml:math> PCA Algorithm Using Artificial Sparsity in Dynamic MRI

Analysis of PCA-based reconstruction method for fault diagnosis

Moving window kernel PCA for adaptive monitoring of nonlinear processes

Developments and Applications of Nonlinear Principal Component Analysis – a Review

Sensor Fault Diagnosis of Nonlinear Processes Based on Structured Kernel Principal Component Analysis

KPCA Based on Feature Samples for Fault Detection

Analysis of Principal Component Analysis-Based Reconstruction Method for Fault Diagnosis

Determining the Optimal Kernel Parameter in KPCA Based on Sample Reconstruction

To Solve Kernel Principal Component Analysis Using Iterative Method

Nonlinear Process Monitoring Using Improved Kernel Principal Component Analysis

Adaptive kernel subspace method for speeding up feature extraction

Learning a Data-Dependent Kernel Function for KPCA-based Nonlinear Process Monitoring

Novel Local Linear Predictor of Chaotic Time Series Based on Kernel Function

Accelerated Kernel Canonical Correlation Analysis with Fault Relevance for Nonlinear Process Fault Isolation

Improved kernel PCA-based monitoring approach for nonlinear processes

An Iterative Algorithm for Robust Kernel Principal Component Analysis

Nonlinear Batch Process Monitoring Using Phase-Based Kernel-Independent Component Analysis-Principal Component Analysis (Kica-Pca)

A Kernel Spatial Complexity-Based Nonlinear Unmixing Method of Hyperspectral Imagery.

Fault detection method based on kernel principal angle