ELM-Randomness-Kernel

ELM-Randomness-Kernel

ID:40354724

大小:734.66 KB

页数:15页

时间:2019-07-31

ELM-Randomness-Kernel_第1页
ELM-Randomness-Kernel_第2页
ELM-Randomness-Kernel_第3页
ELM-Randomness-Kernel_第4页
ELM-Randomness-Kernel_第5页
资源描述:

《ELM-Randomness-Kernel》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、Author'spersonalcopyCognComput(2014)6:376–390DOI10.1007/s12559-014-9255-2AnInsightintoExtremeLearningMachines:RandomNeurons,RandomFeaturesandKernelsGuang-BinHuangReceived:25January2014/Accepted:13March2014/Publishedonline:3April2014ÓSpringerScience+BusinessMediaNewYork2014AbstractExtremelearningmach

2、ines(ELMs)basicallyroleinbothresearchandclassificationrelatedapplicationsgiveanswerstotwofundamentallearningproblems:(1)inthepasttwodecades.SVMachieveshighergeneral-Canfundamentalsoflearning(i.e.,featurelearning,clus-izationperformancethanconventionalartificialneuraltering,regressionandclassification)b

3、emadewithoutnetworksinmostclassificationapplications.Intheoriginaltuninghiddenneurons(includingbiologicalneurons)evenimplementationofSVM,onehastohandleaquadraticwhentheoutputshapesandfunctionmodelingoftheseprogramming(QP)problemwhichisusuallytediousandneuronsareunknown?(2)Doesthereexistunifiedframe-ti

4、meconsuming.AsoneofthemainvariantofSVM,Leastworkforfeedforwardneuralnetworksandfeaturespacesquaresupportvectormachine(LS-SVM)[2]aimstoavoidmethods?ELMsthathavebuiltsometangiblelinkstheQPproblemusingequalityconstraintsinsteadofthebetweenmachinelearningtechniquesandbiologicalinequalityconstraintadopte

5、dinconventionalSVM.Com-learningmechanismshaverecentlyattractedincreasingparedwithSVM,LS-SVMiseaseofimplementation.attentionofresearchersinwidespreadresearchareas.ThisExtremelearningmachines(ELMs)[3–7]becomeattrac-paperprovidesaninsightintoELMsinthreeaspects,viz:tivetomoreandmoreresearchersrecently[8

6、–20].Thisrandomneurons,randomfeaturesandkernels.ThispaperpaperaimstoreviewtheELMfromrandomneuronsandalsoshowsthatintheoryELMs(withthesamekernels)kernelspointofviewandtobuildsomerelationshipandtendtooutperformsupportvectormachineanditsvariantslinksbetweenELM,SVMandotherrelatedmachineinbothregressiona

7、ndclassificationapplicationswithlearningtechniques.Althoughitisoutofquestionthatmucheasierimplementation.SVManditsvariantsachievesurprisingperformanceinmostapplications,differentfromsomecommonconceptin

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。
相关文章
更多
相关标签