中国邮电高校学报(英文) ›› 2012, Vol. 19 ›› Issue (6): 79-88.doi: 10.1016/S1005-8885(11)60321-X

• Others • 上一篇    下一篇

Semi-supervised learning via manifold regularization

毛昱,周延泉   

  1. 北京邮电大学
  • 收稿日期:2012-05-07 修回日期:2012-07-02 出版日期:2012-12-31 发布日期:2012-12-14
  • 通讯作者: 毛昱 E-mail:maomaoyu10@gmail.com
  • 基金资助:

    This work was supported by the Mechanism Socialist Method and Higher Intelligence Theory of the National Natural Science Fund Projects (60873001).

Semi-supervised learning via manifold regularization

  • Received:2012-05-07 Revised:2012-07-02 Online:2012-12-31 Published:2012-12-14
  • Supported by:

    This work was supported by the Mechanism Socialist Method and Higher Intelligence Theory of the National Natural Science Fund Projects (60873001).

摘要:

This paper proposes a novel graph-based transductive learning algorithm based on manifold regularization. First, the manifold regularization was introduced to probabilistic discriminant model for semi-supervised classification task. And then a variation of the expectation maximization (EM) algorithm was derived to solve the optimization problem, which leads to an iterative algorithm. Although our method is developed in probabilistic framework, there is no need to make assumption about the specific form of data distribution. Besides, the crucial updating formula has closed form. This method was evaluated for text categorization on two standard datasets, 20 news group and Reuters-21578. Experiments show that our approach outperforms the state-of-the-art graph-based transductive learning methods.

关键词:

manifold regularization, semi-supervised learning, transductive learning, expectation maximization algorithm, classification, text categorization

Abstract:

This paper proposes a novel graph-based transductive learning algorithm based on manifold regularization. First, the manifold regularization was introduced to probabilistic discriminant model for semi-supervised classification task. And then a variation of the expectation maximization (EM) algorithm was derived to solve the optimization problem, which leads to an iterative algorithm. Although our method is developed in probabilistic framework, there is no need to make assumption about the specific form of data distribution. Besides, the crucial updating formula has closed form. This method was evaluated for text categorization on two standard datasets, 20 news group and Reuters-21578. Experiments show that our approach outperforms the state-of-the-art graph-based transductive learning methods.

Key words:

manifold regularization, semi-supervised learning, transductive learning, expectation maximization algorithm, classification, text categorization