Full-Text Search:
Home|Journal Papers|About CNKI|User Service|FAQ|Contact Us|中文
《Computer Applications and Software》 2010-11
Add to Favorite Get Latest Update

A PARALLEL FEATURE SELECTION ALGORITHM WITH SPECTRAL THEORY

Yang Qinyao Yu Guoxian Lü Le (School of Computer Science and Engineering,South China University of Technology,Guangzhou 510006,Guangdong,China)  
Feature selection with spectral theory(FSST) prefers the features with utmost preservation ability for local information and global discriminative ability.On the basis of analysing FSST via experiment,we parallelized the most time-consuming parts of FSST(normalizing data,construct Laplacian graph and calculate feature score) with divide and conquer strategy,and then proposed a parallel feature selection algorithm with spectral theory(PFSST).Experiments on multicore system proved its effectiveness in parallel.
【CateGory Index】: TP301.6
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