2014 Coupled Group Lasso for Web-Scale CTR Prediction in Display Advertising

2014 Coupled Group Lasso for Web-Scale CTR Prediction in Display Advertising

ID:40378333

大小:553.20 KB

页数:9页

时间:2019-08-01

2014 Coupled Group Lasso for Web-Scale CTR Prediction in Display Advertising_第1页
2014 Coupled Group Lasso for Web-Scale CTR Prediction in Display Advertising_第2页
2014 Coupled Group Lasso for Web-Scale CTR Prediction in Display Advertising_第3页
2014 Coupled Group Lasso for Web-Scale CTR Prediction in Display Advertising_第4页
2014 Coupled Group Lasso for Web-Scale CTR Prediction in Display Advertising_第5页
资源描述:

《2014 Coupled Group Lasso for Web-Scale CTR Prediction in Display Advertising》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、CoupledGroupLassoforWeb-ScaleCTRPredictioninDisplayAdvertisingLingYanYLING0718@SJTU.EDU.CNShanghaiKeyLaboratoryofScalableComputingandSystems,DepartmentofComputerScienceandEngineering,Shang-haiJiaoTongUniversity,ChinaWu-JunLiLIWUJUN@NJU.EDU.CNNationalKeyLaborat

2、oryforNovelSoftwareTechnology,DepartmentofComputerScienceandTechnology,NanjingUniversity,ChinaGui-RongXueGRXUE@ALIBABA-INC.COMAlibabaGroup,ChinaDingyiHanDINGYI.HAN@ALIBABA-INC.COMAlibabaGroup,ChinaAbstract1.IntroductionRecently,onlineadvertisinghasbecomethemos

3、tpopularIndisplayadvertising,clickthroughrate(CTR)andeffectiveapproachtodobrandpromotionandproduc-predictionistheproblemofestimatingtheprob-tmarketing.Itisamulti-billionbusinessonthewebandabilitythatanadvertisement(ad)isclickedwhenaccountsforthemajorityofthein

4、comeforthemajorinter-displayedtoauserinaspecificcontext.Duenetcompanies,suchasGoogle,YahooandAlibaba.Dis-toitseasyimplementationandpromisingperfor-playadvertisingisabigpartofonlineadvertisingwheremance,logisticregression(LR)modelhasbeenadvertiserspaypublishersf

5、orplacinggraphicaladvertise-widelyusedforCTRprediction,especiallyinin-ments(ads)onpublishers’webpages(Chapelleetal.,dustrialsystems.However,itisnoteasyforLR2013).Thepublishersallocatesomepositionsontheirwebtocapturethenonlinearinformation,suchasthepagesandsell

6、themtodifferentadvertisers.Usersvisittheconjunctioninformation,fromuserfeaturesandwebpagesandcanviewthepublishedads.Therearesomeadfeatures.Inthispaper,weproposeanov-otherroles,suchasadagenciesandpublishernetworks,toelmodel,calledcoupledgrouplasso(CGL),forcompo

7、sethecomplexadvertisingsystem(Muthukrishnan,CTRpredictionindisplayadvertising.CGLcan2009).Butthatisnotthefocusofthispaper.Sowewilljustseamlesslyintegratetheconjunctioninformationfocusonthescenarioswithauser-advertiser-publishertri-fromuserfeaturesandadfeatures

8、formodeling.partitebusiness,inwhichthreepartieshaveseparategoalsFurthermore,CGLcanautomaticallyeliminatethatcanbereducedtoaunifiedtaskintheend.Thead-uselessfeaturesforbothusersandad

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。