欢迎来到天天文库
浏览记录
ID:33487414
大小:8.25 MB
页数:67页
时间:2019-02-26
《基于数据挖掘技术的微博好友推荐机制的分析与研究》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、万方数据东北大学硕士学位论文AbstractResearchandImplementationofMicroblogFriends’RecommenderMechanismBasedonDataMiningTechniqueAbstractWiththerapiddevelopmentofthemicrobloggingservices,userscansendandreceivetheirinterestedcontentsbymeansofthisemergingplatformaccordingtotheirownwishesThe
2、microbloghasbecomethemostimportantsocialmediainwhichitsuserscancommunicatewitheachotheraswellasdisseminateinformation.Whenfacingwiththemassofinformationgeneratedfromthemicrobloggingplatformeveryday,howtofindtheusefulinformationhasbecomeacriticalissuethathasdrawnmanyconcer
3、ns.Astheinformationofmicrobiogsisalwaysaccompaniedwiththemicroblogpublishers,findingsimilarmicroblogusershasbecomeanimportantchannelforuserstogettheinformationneeded.Allabovemakemicroblogfriends’recommendationbecomeanewhotspotproblemforbothcommerciaIandresearchcommunities
4、Basedonmicrobloganalysisanddataminingtechnology,thisthesisstudiesonthemechanismofmicroblogfriends’recommendation,whichincludestheconstructionthemodelofmicroblogfriends’recommendationandtheimplementationofrelatedalgorithmsFirstly,theexistingmicroblogfriends’recommendationa
5、lgorithmshavebeendeeplystudied.Relatedtheorieshavebeensummarized,andmultiplefeaturesofmicroblogusershavebeenanalyzedandcombined.Apersonalizedfriends’recommendationmodelthatcanutilizemultiplefeaturesofmicroblogusersincludingcurrentfriendships,personalizedtags,locationinfor
6、mationandthehottesttopicshasbeenconstructed.Secondly,basedonthemodelconstructed,amicroblogfriends’recommendationalgorithmisdesignedandimplemented.Specially,afriends’recommendationprocessisdesignedatfirst,Then,theprocessofcandidatefriends’generationandthesimilaritycalculat
7、ionofusertags,locationinformation,andinterestedhottopicsaredesignedrespectively.Finally,apersonalizedfriends’recommendationalgorithmcombiningmultiplefeaturesofmicrobloguserisproposedLastly,inordertodemonstratethefeasibilityandadvantagesoftheproposedmodelandalgorithms,thec
8、ross—validationmethodisemployedtoverifytheeffectivenessofthe.V一万方数据东北大学硕士学位论文Abstractproposedper
此文档下载收益归作者所有