Quasi-newton method for L_P multiple kernel learning

详细信息    查看全文

• 作者：Hu Qinghui^a ; ^bAuthor Vitae ; Wei Shiwei^b ; ^{swwei_2001@163.com" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Li Zhiyuan^bAuthor Vitae ; Liu Xiaogang^aAuthor Vitae
• 关键词：Multiple kernel learning ; Quasi-Newton method ; Alternating optimization
• 刊名：Neurocomputing
• 出版年：2016
• 出版时间：19 June 2016
• 年：2016
• 卷：194
• 期：Complete
• 页码：218-226
• 全文大小：655 K

文摘

Multiple kernel learning method has more advantages over the single one on the model’s interpretability and generalization performance. The existing multiple kernel learning methods usually solve SVM in the dual which is equivalent to the primal optimization. Research shows solving in the primal achieves faster convergence rate than solving in the dual. This paper provides a novel L_P-norm(P>1) constraint non-spare multiple kernel learning method which optimizes the objective function in the primal. Subgradient and Quasi-Newton approach are used to solve standard SVM which possesses superlinear convergence property and acquires inverse Hessian without computing a second derivative, leading to a preferable convergence speed. Alternating optimization method is used to solve SVM and to learn the base kernel weights. Experiments show that the proposed algorithm converges rapidly and that its efficiency compares favorably to other multiple kernel learning algorithms.