The optimization of the kind and parameters of kernel function in KPCA for process monitoring

Abstract Kernel principal component analysis (KPCA) has been widely used in chemical processes monitoring due to its simple principle. However, how to select the kind and parameters of kernel function still limits the application of the method until now. In this paper, an optimization method based on genetic algorithm is developed to choose proper kind and parameters of kernel function. In this method, kernel kind and parameters are seen as decision variables of optimization, using correct monitoring rate, number of principal components, and statistical control limit of square prediction error (SPE) as multi-objective. For this specific problem, the fitness function, the algorithm of genetic selection, crossover and mutation are designed to ensure the diversity of kernel function and more selected chances of optimal individual in evolution process. A simple example and penicillin fermentation process are used to investigate the potential application of the proposed method; simulation results show that the proposed method is effective.

[1]  Massimiliano Pontil,et al.  Support Vector Machines for 3D Object Recognition , 1998, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  Julian Morris,et al.  Nonlinear multiscale modelling for fault detection and identification , 2008 .

[3]  C. I. Mosier.  I. Problems and Designs of Cross-Validation 1 , 1951 .

[4]  Yingwei Zhang,et al.  Enhanced statistical analysis of nonlinear processes using KPCA, KICA and SVM , 2009 .

[5]  Zhiqiang Ge,et al.  Improved kernel PCA-based monitoring approach for nonlinear processes , 2009 .

[6]  Steven M. LaValle,et al.  On the Relationship between Classical Grid Search and Probabilistic Roadmaps , 2004, Int. J. Robotics Res..

[7]  S. Wold.  Cross-Validatory Estimation of the Number of Components in Factor and Principal Components Models , 1978 .

[8]  Shufeng Tan,et al.  Reducing data dimensionality through optimizing neural network inputs , 1995 .

[9]  In-Beum Lee,et al.  Fault identification for process monitoring using kernel principal component analysis , 2005 .

[10]  Cheng-Lung Huang,et al.  A GA-based feature selection and parameters optimizationfor support vector machines , 2006, Expert Syst. Appl..

[11]  Chih-Jen Lin,et al.  A Simple Decomposition Method for Support Vector Machines , 2002, Machine Learning.

[12]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[13]  U. Kruger,et al.  Moving window kernel PCA for adaptive monitoring of nonlinear processes , 2009 .

[14]  Bernhard Schölkopf,et al.  Nonlinear Component Analysis as a Kernel Eigenvalue Problem , 1998, Neural Computation.

[15]  Gang Rong,et al.  Nonlinear process monitoring based on maximum variance unfolding projections , 2009, Expert Syst. Appl..

[16]  T. McAvoy,et al.  Nonlinear principal component analysis—Based on principal curves and neural networks , 1996 .

[17]  Ivan Dvořák,et al.  Singular-value decomposition in attractor reconstruction: pitfalls and precautions , 1992 .

[18]  Nello Cristianini,et al.  An Introduction to Support Vector Machines and Other Kernel-based Learning Methods , 2000 .

[19]  Liangsheng Qu,et al.  Evolving kernel principal component analysis for fault diagnosis , 2007, Comput. Ind. Eng..

[20]  C. Yoo,et al.  Nonlinear process monitoring using kernel principal component analysis , 2004 .

[21]  Nagiza F. Samatova,et al.  An SVM-based algorithm for identification of photosynthesis-specific genome features , 2003, Computational Systems Bioinformatics. CSB2003. Proceedings of the 2003 IEEE Bioinformatics Conference. CSB2003.

[22]  S. Qin,et al.  Selection of the Number of Principal Components: The Variance of the Reconstruction Error Criterion with a Comparison to Other Methods† , 1999 .

[23]  M. Kramer.  Nonlinear principal component analysis using autoassociative neural networks , 1991 .

[24]  In-Beum Lee,et al.  Nonlinear dynamic process monitoring based on dynamic kernel PCA , 2004 .

[25]  Ricardo Massa Ferreira Lima,et al.  GA-based method for feature selection and parameters optimization for machine learning regression applied to software effort estimation , 2010, Inf. Softw. Technol..

[26]  Gunnar Rätsch,et al.  Kernel PCA and De-Noising in Feature Spaces , 1998, NIPS.

[27]  Engin Avci.  Selecting of the optimal feature subset and kernel parameters in digital modulation classification by using hybrid genetic algorithm-support vector machines: HGASVM , 2009, Expert Syst. Appl..

[28]  Thorsten Joachims,et al.  Text categorization with support vector machines , 1999 .

[29]  Mingtian Zhou,et al.  Feature selection and parameter optimization for support vector machines: A new approach based on genetic algorithm with feature chromosomes , 2011, Expert Syst. Appl..

[30]  Jin Hyun Park,et al.  Fault detection and identification of nonlinear processes based on kernel PCA , 2005 .

[31]  Ryo Saegusa,et al.  Nonlinear principal component analysis to preserve the order of principal components , 2003, Neurocomputing.

[32]  Wojtek J. Krzanowski,et al.  Cross-Validation in Principal Component Analysis , 1987 .

[33]  Gunnar Rätsch,et al.  Input space versus feature space in kernel-based methods , 1999, IEEE Trans. Neural Networks.

[34]  Zhao Chun-hui.  New Nonlinear Principal Analysis Method Based on RBF Neural Network , 2007 .

[35]  ChangKyoo Yoo,et al.  Fault detection of batch processes using multiway kernel principal component analysis , 2004, Comput. Chem. Eng..

[36]  Paul Geladi,et al.  Principal Component Analysis , 1987, Comprehensive Chemometrics.

[37]  Chih-Hung Wu,et al.  A Novel hybrid genetic algorithm for kernel function and parameter optimization in support vector regression , 2009, Expert Syst. Appl..