Full-Text Search:
Home|Journal Papers|About CNKI|User Service|FAQ|Contact Us|中文
《Journal of Hunan University(Natural Sciences)》 2008-07
Add to Favorite Get Latest Update

On Parallel Learning Based on Support Vector Machines

JIN Xiao-ming1,2,WEN Yi-min3 (1.Business School,Central South Univ,Changsha,Hunan 410083,China;2.Hunan Institute of Technology,Hengyang,Hunan 421002,China;3.College of Electrical and Information Engineering,Hunan Univ,Changsha,Hunan 410082,China)  
This paper has proposed a parallel combine SVMs on the basis of two information fusion rules(B-SVMs).B-SVMs randomly decomposes a large-scale task into many smaller and simpler sub-tasks in the training phase and uses the two information fusion rules to make decisions for final classification in the test phase.B-SVMs has been compared with single SVMs that was trained on entire training data set,parallel SVMs combined by majority voting(MV-SVMs),and one kind of fast modular SVMs(FM-SVMs).Experiment results on four problems have shown that B-SVMs can get a higher accuracy than MV-SVMs and FM-SVMs.The proposed algorithm can significantly reduce training and test time.More importantly, it produces a test accuracy that is almost the same as single SVMs does.
【Fund】: 国家自然科学基金资助项目(70540014)
【CateGory Index】: TP391.41
Download(CAJ format) Download(PDF format)
CAJViewer7.0 supports all the CNKI file formats; AdobeReader only supports the PDF format.
©2006 Tsinghua Tongfang Knowledge Network Technology Co., Ltd.(Beijing)(TTKN) All rights reserved