Large sample classification based on Dual-SVM
HU Xiao-sheng
Abstract：The dual support vector machine algorithm is proposed to improve the learning efficiency for large-scale sample classification based on two-phase training processes. In the first step, K-means clustering is performed to original training data of each class, all clustering centers are extracted and made up as a reduced-size training set for the first SVM training, then some cluster centers obtained as support vectors are regarded to be located close to the hyperplane border. In the second step, original samples are contained in clusters for which the cluster centers are obtained to be close to the hyperplane border, these border samples are then used in the second SVM classifier. Experimental results show that the proposed method leads to much faster SVM training without reducing the classification accuracy.
Keywords：support vector machinelarge-scale classificationclusteringsample selection
Publication Date：2015-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：5( 26-30 )