Improvement on multivariate statistical process monitoring using multi-scale ICA

Authors:
Fei Liu;Chang-Ying Wu
Affiliations:
Institute of Automation, Southern Yangtze University, Wuxi, P.R. China;Institute of Automation, Southern Yangtze University, Wuxi, P.R. China
Venue:
ICA'06 Proceedings of the 6th international conference on Independent Component Analysis and Blind Signal Separation
Year:
2006

Citing 8
Cited 1

A Theory for Multiresolution Signal Decomposition: The Wavelet Representation

IEEE Transactions on Pattern Analysis and Machine Intelligence
Elements of information theory

Elements of information theory
Ten lectures on wavelets

Ten lectures on wavelets
Independent component analysis: algorithms and applications

Neural Networks
Multivariate statistical process control with industrial applications

Multivariate statistical process control with industrial applications
Independent Component Analysis: Principles and Practice

Independent Component Analysis: Principles and Practice
A multiscale framework for blind separation of linearly mixed signals

The Journal of Machine Learning Research
Fast and robust fixed-point algorithms for independent component analysis

IEEE Transactions on Neural Networks

Variable window adaptive Kernel Principal Component Analysis for nonlinear nonstationary process monitoring

Computers and Industrial Engineering

Quantified Score

Hi-index	0.00

Visualization

Abstract

A multi-scale independent component analysis (ICA) approach is investigated for industrial process monitoring. By integrating the ability of wavelet on multi-scale analysis and that of ICA on extracting independent components for non-Gaussian process variables, the multivariate statistical monitoring techniques can obtain improved performance. Contrastive tests have been carried out on the famous benchmark chemical plant among ICA-like and PCA-like methods, which reveals that multi-scale ICA approach has lower missed detection rate of faults.