Title :
A Parameters Selection Method of SVM
Author :
Kou, Deqi ; Zhang, Yuan ; Zheng, Hanyue
Author_Institution :
Dept. of Tech. Support Eng., Acad. of Armored Forces Eng., Beijing, China
Abstract :
An improved artificial fish swarm algorithm called ASFSA is proposed. It could facilitate the selection of values for Step and Visual to meet the balance of algorithm speed and effect. A new SVM parameters selection method based on the ASFSA is described, and the kernel parameter γ and regularization parameter C can both be optimized well. The application case shows that the performance of SVM with optimized parameters is good, so the method is feasible and effective.
Keywords :
algorithm theory; support vector machines; ASFSA; SVM parameter selection method; artificial fish swarm algorithm; kernel parameter; regularization parameter; support vector machines; Kernel; Marine animals; Optimization; Support vector machines; Testing; Training; Visualization;
Conference_Titel :
Computational Intelligence and Software Engineering (CiSE), 2010 International Conference on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4244-5391-7
Electronic_ISBN :
978-1-4244-5392-4
DOI :
10.1109/CISE.2010.5676994
