The combining kernel PCA with PSO-SVM for chaotic time series prediction model

• Published in: 2009 International Conference on Machine Learning and Cybernetics Vol. 1; pp. 467 - 472
• Main Authors: Qi-Song Chen, Xin Zhang, Shi-Huan Xiong, Xiao-Wei Chen
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.07.2009
• Subjects: Chaos; Feature extraction; Kernel; KPCA; Machine learning; Pattern recognition; Prediction model; Predictive models; Principal component analysis; PSO; Support vector machine classification; Support vector machines; SVM; Time series analysis
• DOI: 10.1109/ICMLC.2009.5212558
• ISBN: 9781424437023; 1424437024
• ISSN: 2160-133X

Chaotic time series analysis or forecasting is an important and complex problem in machine learning. As an effective tool, support vector machine (SVM) has been broadly adopted in pattern recognition and machine learning fields. In developing a successful SVM classifier, eliminating noise and extracting feature are very important. This paper proposes the application of kernel PCA to LS-SVM for feature extraction. Then PSO algorithm is employed to optimization of these parameters in LS-SVM. The novel chaotic time series analysis model integrates the advantages of wavelet transform, KPCA, PSO and LS-SVM. Compared with other predictors, this model has greater generality ability and higher accuracy.