首页 | 本学科首页   官方微博 | 高级检索  
     

半监督学习的Co-training算法研究
引用本文:刘蓉.半监督学习的Co-training算法研究[J].电脑编程技巧与维护,2010(14):4-5.
作者姓名:刘蓉
作者单位:长沙医学院计算机系,长沙,410219
基金项目:湖南省教育厅科学研究项目 
摘    要:介绍一种基于半监督学习的协同训练(Co-training)分类算法,当可用的训练样本比较少时,使用传统的方法进行分类,如决策树分类,将无法得到用户满意的结果,而且它们需要大量的标记样本。事实上,获取有标签的样本的代价是相当昂贵的。于是,使用较少的已标记样本和大量的无标记样本进行协同训练的半监督学习,成为研究者首选。

关 键 词:半监督学习  协同训练(Co-training)  分类

The Research of Co-training Algorithm Based on Semi-supervised Learning
LIU Rong.The Research of Co-training Algorithm Based on Semi-supervised Learning[J].Computer Programming Skills & Maintenance,2010(14):4-5.
Authors:LIU Rong
Affiliation:LIU Rong (Changsha The medical school the computer department,Changsha 410219)
Abstract:This paper presents a semi-supervised learning based on collaborative training (Co-training) classification algorithm,when the comparison of available training samples is low,the use of traditional classification methods,such as decision tree,the user will not get satisfactory results.And they need a large number of labeled samples,can be a matter of fact,to obtain a sample of the price tag is quite expensive.So,with less marked samples and unmarked samples a large number of collaborative training of semi-supervised learning,a researcher of choice.
Keywords:Semi-supervised learning  Collaborative training  Category
本文献已被 维普 万方数据 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