A Bootstrapping Approach to Identifying Relevant Tweets for Social TV

A Bootstrapping Approach to Identifying Relevant Tweets for Social TV

ID:40022705

大小:182.24 KB

页数:4页

时间:2019-07-17

A Bootstrapping Approach to Identifying Relevant Tweets for Social TV_第1页
A Bootstrapping Approach to Identifying Relevant Tweets for Social TV_第2页
A Bootstrapping Approach to Identifying Relevant Tweets for Social TV_第3页
A Bootstrapping Approach to Identifying Relevant Tweets for Social TV_第4页
资源描述:

《A Bootstrapping Approach to Identifying Relevant Tweets for Social TV》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、ABootstrappingApproachtoIdentifyingRelevantTweetsforSocialTVOvidiuDan∗JunlanFengBrianD.DavisonLehighUniversityAT&TLabsResearchLehighUniversityBethlehem,PA,USAFlorhamPark,NJ,USABethlehem,PA,USAovd209@lehigh.edujunlan@research.att.comdavison@cse.lehigh.eduAbstractbootstrappin

2、gapproachtotrainingaclassifierusingalim-itedamountoftrainingdataandalargedatasetofunla-ManufacturersofTVsetshaverecentlystartedaddingsocialbeledmessages.Theclassifierisdesignedtogeneralizetomediafeaturestotheirproducts.Someoftheseproductsdis-TVshowsthatwerenotpartofthetrainin

3、gset.Second,playmicrobloggingmessagesrelevanttotheTVshowwhichbasedonourintuitionandexperiments,wedevelopfeaturestheuseriscurrentlywatching.However,suchsystemssufferfromlowprecisionandrecallwhentheyusethetitleofthewhichcanbeusedtofiltermessagesforSocialTVappli-showtosearchfor

4、relevantmessages.Titlesofsomepopularcations.Third,weevaluateourbootstrappedclassifieronashowssuchasLostorSurvivorarehighlyambiguous,result-combinationoftwomanuallylabeleddatasets.Weshowthatinginmessagesunrelatedtotheshow.Thus,thereisaneedtoourclassifierachieveshighprecision,s

5、uccessfullygeneral-developfilteringalgorithmsthatcanachievebothhighpreci-izestounseentelevisionshows,andmatchesorsurpassessionandrecall.FilteringmicrobloggingmessagesforSocialthebaselinewhichistrainedonspecifictelevisionshows.TVposesseveralchallenges,includinglackoftrainingda

6、ta,OnlinesocialnetworkssuchasTwitterhaveattractedlackofpropergrammarandcapitalization,lackofcontextmuchinterestfromtheacademiccommunityinthelastfewduetotextsparsity,etc.years(Kwaketal.2010).SocialTVprojectshaveusedau-Wedescribeabootstrappingalgorithmwhichusesasmalldio,video

7、,andtextchatlinkstotestinteractionbetweenmanuallylabeleddataset,alargedatasetofunlabeledmes-userswatchingTVinseparaterooms.Morerecentlytheresages,andsomedomainknowledgetoderiveahighprecisionhasbeenworkoncombiningthesetwofieldsbydisplay-classifierthatcansuccessfullyfiltermicrob

8、loggingmessagesingmessagesfromsocialnetworksinSocialTVinterfaceswhichdiscusstelevi

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。