Title :
Least Square Support Vector Machine Based on Improved Particle Swarm Optimization to Short-term Forecasting
Author :
Zhang, Dabin ; Peng, Sen ; Duan, Yuting ; Zhang, Wensheng
Abstract :
Forecasting based on least squares support vector machine (LS-SVM) method can be a very good track historical data, and there have a good predictive ability of extrapolation. However, parameter selection is an import work in the application of LS-SVM as it is related to the performance of the constructed predicting. Therefore, an improved particle swarm optimization (IPSO) algorithm was proposed to optimize parameters selection, IPSO for selecting the global optimum parameters of LS-SVM automatically, and avoiding the defects of premature convergence of PSO algorithm. The empirical results show that the improved approach has a better performance and is more effective than other approaches.
Keywords :
extrapolation; forecasting theory; particle swarm optimisation; support vector machines; extrapolation; improved particle swarm optimization; least square support vector machine; parameter selection; short-term forecasting; Convergence; Educational institutions; Forecasting; Kernel; Optimization; Particle swarm optimization; Support vector machines; Empirical; Forecasting; IPSO; LS-SVM; Parameter;
Conference_Titel :
Business Intelligence and Financial Engineering (BIFE), 2011 Fourth International Conference on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4577-1541-9
DOI :
10.1109/BIFE.2011.76