Convergence rate of the semi-supervised greedy algorithm

详细信息    查看全文

• 作者：Hong Chen^a ; ^b ; ^{chenhongml@163.com} ; ^{chenh@mail.hzau.edu.cn} ; Yicong Zhou^b ; Yuan Yan Tang^b ; Luoqing Li^c ; Zhibin Pan^a
• 关键词：Semi-supervised learning ; Sparse ; Greedy algorithm ; Data-dependent hypothesis space ; Generalization error
• 刊名：Neural Networks
• 出版年：2013
• 出版时间：August, 2013
• 年：2013
• 卷：44
• 期：Complete
• 页码：44-50
• 全文大小：313 K

文摘

This paper proposes a new greedy algorithm combining the semi-supervised learning and the sparse representation with the data-dependent hypothesis spaces. The proposed greedy algorithm is able to use a small portion of the labeled and unlabeled data to represent the target function, and to efficiently reduce the computational burden of the semi-supervised learning. We establish the estimation of the generalization error based on the empirical covering numbers. A detailed analysis shows that the error has decay. Our theoretical result illustrates that the unlabeled data is useful to improve the learning performance under mild conditions.