首页 | 本学科首页   官方微博 | 高级检索  
     

基于Tri-training的主动学习算法
引用本文:张雁,吴保国,吕丹桔,林英.基于Tri-training的主动学习算法[J].计算机工程,2014(6):215-218,229.
作者姓名:张雁  吴保国  吕丹桔  林英
作者单位:[1]北京林业大学信息学院,北京100083 [2]西南林业大学计算机与信息学院,昆明650224 [3]云南大学软件学院,昆明650091
基金项目:云南省教育厅科研基金资助项目(2010Y290,2012C098).
摘    要:半监督学习和主动学习都是利用未标记数据,在少量标记数据代价下同时提高监督学习识别性能的有效方法。为此,结合主动学习方法与半监督学习的Tri-training算法,提出一种新的分类算法,通过熵优先采样算法选择主动学习的样本。针对UCI数据集和遥感数据,在不同标记训练样本比例下进行实验,结果表明,该算法在标记样本数较少的情况下能取得较好的效果。将主动学习与Tri-training算法相结合,是提高分类性能和泛化性的有效途径。

关 键 词:半监督学习  主动学习  Tri-training算法  熵优先采样  Tri-EPS算法

Active Learning Algorithm Based on Tri-training
ZHANG Yan,WU Bao-guo,LV Dan-ju,LIN Ying.Active Learning Algorithm Based on Tri-training[J].Computer Engineering,2014(6):215-218,229.
Authors:ZHANG Yan  WU Bao-guo  LV Dan-ju  LIN Ying
Affiliation:1 School of Information, Beijing Forestry University, Beijing 100083, China ;2. School of Computer and Information, Southwest Forestry University, Kunming 650224, China; 3. School of Software, Yunnan University, Kunming 650091, China)
Abstract:Both semi-supervised learning and active learning attempt to exploit the unlabeled data to improve the recognition rate of supervised learning algorithms and minimize the cost of data labeling. So this paper proposes an algorithm to select samples in active learning such as Entropy Priority Sampling(EPS). It combines with the Tri-training algorithm and active learning method. Experimental results on both the UCI and image datasets under different proportion of marker training samples show that, this algorithm can obtain better result in the case of fewer labeled examples, and the combination of the active learning with semi-supervised learning is an effective way to improve the performance and generalization.
Keywords:semi-supervised learning  active learning  Tri-training algorithm  Entropy Priority Sampling(EPS)  Tri-EPS algorithm
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号