Selective ensemble of SVDDs with Renyi entropy based diversity measure
详细信息 查看全文
- 作者：Hong-Jie Xinga ; hjxing@hbu.edu.cn" class="auth_mail" title="E-mail the corresponding authorAuthor Vitae ; Xi-Zhao WangbAuthor Vitae
- 关键词：One-class classification ; Selective ensemble ; Support vector data description ; Renyi entropy
- 刊名：Pattern Recognition
- 出版年：2017
- 出版时间：January 2017
- 年：2017
- 卷：61
- 期：Complete
- 页码：185-196
- 全文大小：1170 K
文摘
- •
A selective ensemble strategy for support vector data descriptions is proposed.
- •
The definition of the radius of ensemble is given.
- •
The Renyi entropy based diversity measure is defined.
- •
An l1-norm based regularization term is introduced to fulfill selective ensemble.
- •
The optimization problem is solved by the half-quadratic optimization technique.