資源描述：
《基于統(tǒng)計(jì)的個(gè)性化微博信息與用戶推薦》由會(huì)員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫。
1、哈爾濱工業(yè)大學(xué)工學(xué)碩士學(xué)位論文AbstractInrecentyears,astheInternettoflourish,especiallyaccompaniedbytheriseofsocialnetworkswebsites,itwasdiscoveredthattheInternetbegantoappearonthephenomenonofinformationoverload,toomuchinformationwillnothelppeoplemoreeasilytofindinformation,butmoredifficult,fromt
2、helargeamountsofinformation,peoplecannotfindoutwhatisimportantandwhatisoptional,andsocialnetworkingisaself-medianetworkapplications,anyone,atanytimecanpostatweetonthewebsite,apparentlytoincreasethedegreeofinformationoverload.BasedonstatisticallearningpersonalizedRecommendationinSo
3、cialMediausestatisticalmachinelearningmethodtoestablishpersonalizedusermodel,tohelpusersavoidinformationoverload,themodeldiscoveryimportanttweetandfriendformicrobloguser,thistechnologyisalsoveryimportantforsocialnetworkwebsitestoimproveuserexperience,atthesametime,therecommendersy
4、stemandsocialnetworkingisthehotresearchfieldcurrently,webelievethattherecommendationinthesocialnetworkingmediaisalsoimportant.Themaincontentisdividedintotweetrecommendationandfriendrecommendationtwoparts.Fortweetsrecommendation,weproposeapersonalizedmethodtopushthetextcontenttothe
5、user,wegivethefeatureanalysisandexperimentsshowtheeffectivenessofourmethod.Forfriendsrecommendation,wehopethatbyfriendsrecommendedcarefullyforusersofsocialmedia,theywillbeabletochoosetherightfriendsandthusabletoachievetheeffectofinformationfilter.Weproposedseveralmethodsbasedoncol
6、laborativefiltering,heuristic,linkprediction,topicmodel.Finally,inordertoboostthepredictionaccuracyoffriendsrecommendation,andtotakeadvantageofmulti-models,wecombineallofmodelsbyusingensemblelearningtechnology.Theexperimentsprovedtheeffectivenessofourproposedmethods,fortheinformat
7、ionrecommended,wegettherecallrate0.49,fortheuserrecommendation,weexaminedtheperformanceofavarietyofrecommendationalgorithms,andboostperformancethroughtheensemblelearning.-II-哈爾濱工業(yè)大學(xué)工學(xué)碩士學(xué)位論文KeywordsRecommenderSystem,SocialNetwork,CollaborativeFiltering,EnsembleLearning-III-哈爾濱工業(yè)大學(xué)工
8、學(xué)碩士學(xué)位論文目錄摘要......................
