全文下载：2013050528
张志锋， 范乃梅
(郑州轻工业学院 软件学院, 河南 郑州 450002)
摘要： 针对当前主动学习策略直接用于SVM（Support Vector Machine）分类器时存在的泛化能力不强的问题，提出的两层协同主动学习策略TLCALS(Two-Level Collaboration Active Learning Strategy)应用了协同训练的思想，能深层挖掘未标记样本数据的分布知识。实验表明，TLCALS策略能够合理指定TSVM(Transducive Support Vector Machine)算法中的正样本数，在典型指标测试中都表现出了一定的优越性。
关键词： 支持向量机； 两层协同主动学习策略； 贝叶斯网络
中图分类号： TP 315文献标志码： A
Classification Algorithm Based on Two-Level Collaboration Active Learning Strategy
ZHANG Zhi-feng, FAN Nai-mei
(SchoolofSoftware,ZhengzhouUniversityof Light Industry,Zhengzhou450002,China)
Abstract: To deal with the poor generalization problem when active learning strategy is directly used in SVM classifier, a TLCALS (Two-Level Collaboration Active Learning Strategy) is proposed. It uses the idea of co-training, and can deeply mine the distribution knowledge. The experiment results show that TLCALS strategy can determine the positive labeled sample numbers reasonable and demonstrate its superiority in typical indicator tests.
Key words:support vector machine; two-level collaboration active learning strategy; Bayesian network
收稿日期：2012-11-12
作者简介： 张志锋(1978—)，男，讲师.
