Semi-supervised support vector classification with self-constructed Universum

详细信息    查看全文

• 作者：Yingjie Tian^a ; ^b ; ^{tyj@ucas.ac.cn" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Ying Zhang^c ; ^{zhangying112@mails.ucas.ac.cn" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Dalian Liu^d ; ^{ldlluck@sina.com" class="auth_mail" title="E-mail the corresponding author}Author Vitae
• 关键词：Semi-supervised ; Classification ; Universum ; Support vector machine
• 刊名：Neurocomputing
• 出版年：2016
• 出版时间：12 May 2016
• 年：2016
• 卷：189
• 期：Complete
• 页码：33-42
• 全文大小：1676 K

文摘

In this paper, we propose a strategy dealing with the semi-supervised classification problem, in which the support vector machine with self-constructed Universum is iteratively solved. Universum data, which do not belong to either class of interest, have been illustrated to encode some prior knowledge by representing meaningful concepts in the same domain as the problem at hand. Our new method is applied to seek more reliable positive and negative examples from the unlabeled dataset step by step, and the Universum support vector machine(U$\mathfrak{U}$-SVM) is used iteratively. Different Universum data will result in different performance, so several effective approaches are explored to construct Universum datasets. Experimental results demonstrate that appropriately constructed Universum will improve the accuracy and reduce the number of iterations.