计算机应用研究2017，Vol.34Issue(2)：351-356,6.DOI:10.3969/j.issn.1001-3695.2017.02.007
基于域相关性与流形约束的多源域迁移学习分类算法
Multi-source transfer classification learning based on combination of domain relevance and manifold constraint
摘要
Abstract
In many traditional machine learning algorithms,a major assumption was that the training samples and the test samples had the same distribution.However,this assumption did not hold in many real applications.In recent years,transfer learning had attracted a significant amount of attention to solve this problem.The relationship between domains affected the effectiveness of the transfer.Rather than improving the learning,brute force leveraging of a source poorly related to the target might decrease the classifier performance,i.e.,negative transfer.This paper proposed a novel multi-source transfer learning method based on multi-similarity.The method explored more accurate relationship between the source and target domain by multi-similarity metric.Then,the method transferred the knowledge of the sources to the target based on smoothness assumption,which enforced that the target classifier shared similar decision values with the relevant source classifiers on the unlabeled instances from the target domain.Experimental results on toy and real-life datasets demonstrate that the proposed method can increase the chance of finding the sources closely related to the target to reduce the negative transfer and also imports more knowledge from multiple sources for the target learning.关键词
迁移学习/多源域迁移/域相似性/流形假设Key words
transfer learning/multiple source transfer/domain similarity/manifold assumption分类
信息技术与安全科学引用本文复制引用
刘振,杨俊安,刘辉,王伟..基于域相关性与流形约束的多源域迁移学习分类算法[J].计算机应用研究,2017,34(2):351-356,6.基金项目
国家“863”计划资助项目 （）
安徽省自然科学基金资助项目(1308085QF99,1408085MKL46) （1308085QF99,1408085MKL46）
