Title :
Time Series Prediction Based on Support Vector Machines Experts and Genetic Algorithm
Author_Institution :
China Univ. of Pet., Dongying
Abstract :
A new time series prediction method based on support vector machines (SVMs) experts and genetic algorithm (GA) is proposed. The proposed method has a three-stage architecture. In the first stage, self-organizing feature map (SOM) is used as a clustering algorithm to partition the whole input space into several disjointed regions. Then, in the second stage, GA is adopted to determine the parameter combination of the SVMs expert corresponding to the partitioned region obtained above. In the last stage, the different SVMs experts in the different input-output spaces are constructed and used to predict time series. The simulation result shows that the multiple SVMs experts achieve significant improvement in the generalization performance in comparison with the single SVMs model. In addition, introducing GA to solve the parameter combination involved in SVMs expert avoids determining the parameters from experience.
Keywords :
genetic algorithms; pattern clustering; prediction theory; self-organising feature maps; support vector machines; time series; clustering algorithm; genetic algorithm; self-organizing feature map; support vector machines experts; time series prediction; Clustering algorithms; Control engineering; Genetic algorithms; Neural networks; Partitioning algorithms; Petroleum; Prediction methods; Quadratic programming; Signal processing algorithms; Support vector machines;
Conference_Titel :
Fuzzy Systems and Knowledge Discovery, 2007. FSKD 2007. Fourth International Conference on
Conference_Location :
Haikou
Print_ISBN :
978-0-7695-2874-8
DOI :
10.1109/FSKD.2007.590