国外计算机科学类论文,推荐系统
302J.K.Kimetal./ElectronicCommerceResearchandApplications1 (2002) 301–313
recommendationtechnique,whichhasbeenusedinanumberofdifferentapplicationssuchasrecom-mendingWebpages,movies,articlesandproducts.CFtechniquesidentifycustomers(neighbors)whoseinterestsaresimilartothoseofagivencustomerandrecommendproductstheyprefer.However,despitetheirsuccess,theirwidespreadusehasexposedsomewell-knownlimitations,suchasearlyrater,graysheep,sparsityandscalability,whichcanleadtopoorrecommendations[9,27].
RecentresearchhassuggestedWebusageminingasanenablertoovercometheproblemsassociatedwithCFsinceitwillreducetheneedforobtainingsubjectiveuserratings[23].Theclickstreaminonlinestoresprovidesinformationessentialtounder-standingtheshoppingorprepurchasebehaviorsofcustomerssuchaswhatproductstheysee,whatproductstheyaddtotheshoppingcart,andwhatproductstheybuy.Webusageminingcanmakeamoreaccurateanalysisofcustomer’spreferencesacrossallproductsthansubjectiveuserratings.Theclickstreamcanalsobeusedtoidentifyavarietyofrelationsbetweenproductsfromcustomer’sshoppingbehavior.
Thequalityoftherecommendationshasanim-portanteffectonthecustomer’sfutureshoppingbehavior.Unquali edrecommendationsmaycausetwotypesofcharacteristicerrors:falsenegativeandfalsepositive[27].Theformerreferstheproductthatisnotrecommended,althoughthecustomerwouldpreferit,andthelatterreferstheproductthatisrecommended,althoughthecustomerwouldnotpreferit.Inane-commerceenvironment,themostimportanterrortoavoidisthefalsepositive,becausethiserrorwillleadtoangrycustomersandthustheywillbeunlikelytorevisitthesite.
Inthispaper,weproposeanewpersonalizedrecommendationprocedure,WebCF-DT(WebusageminingdrivenCollaborativeFiltering-basedrecom-mendationprocedureusingDecisionTree).Decisiontree-basedpredictiveminingisadoptedtominimizefalsepositivesbymakingrecommendationsnotforallthecustomersbutforselectedcustomerswhoaremostlikelytobuyrecommendedproducts.WebCF-DTwasappliedtoarealInternetshoppingmallforevaluationoftheprocedure.Someexperimentresultsareprovided.
WebeginbyreviewingpreviousworksrelatedtoourresearchinSection2.Wede nerecommendation
problemstoclarifyourobjectivesinSection3.InSection4,thesuggestedprocedureisexplained.Theperformanceevaluationfortheprocedureisdis-cussedinSection5.Finally,wesummarizeourcontributionswithsuggestionsforfutureresearchareasinSection6.
2 .Relatedworks
2 .1.Webusagemining
WebusageminingistheprocessofapplyingdataminingtechniquestothediscoveryofbehaviorpatternsbasedonWebdata.Withtheadvanceofe-commerce,theimportanceofWebusageminingisgrowinglargerthanbefore.TheoverallprocessofWebusageminingisgenerallydividedintotwomaintasks:datapreparationandpatterndiscovery.Thedatapreparationtasksbuildaserversession lewhereeachsessionisasequenceofrequestsofdifferenttypesmadebyasingleuserduringasinglevisittoasite.Coolyetal.[10]presentedadetaileddescriptionofdatapreparationmethodsforminingWebbrowsingpatterns.Thepatterndiscoverytasksinvolvethediscoveryofassociationrules,sequentialpatterns,usageclusters,pageclusters,userclassi ca-tionsoranyotherpatterndiscoverymethod[22].TheusagepatternextractedfromWebdatacanbeappliedtoawiderangeofapplicationssuchasWebpersonalization,systemimprovement,sitemodi ca-tion,businessintelligencediscovery,usagecharac-terization,etc.[32].
ThisresearchrecommendsproductsbasedonWebusagedataaswellasproductpurchasedataandcustomer-relateddata.Therehavebeenseveralcus-tomerbehaviormodelsfore-commerce,whichhave
differentanalysispurposes.Menasce
´etal.[21]presentedastatetransitiongraph,calledaCustomerBehaviorModelGraph(CBMG),thatisusedtodescribethebehaviorofgroupsofcustomerswhoexhibitsimilarnavigationalpatterns.VandeMeer,DuttaandDatta[33]developedausernavigationmodeldesignedforsupportingandtrackingdynamicuserbehaviorinonlinepersonalization.Themodelsupportsthenotionofaproductcatalog,usernaviga-tionoverthiscataloganddynamiccontentdelivery.LeeandPodlaseck[19]providedadetailedcasestudyoftheanalysisofclickstreamdatafroman
Research on personalized recommendation system in ...INTRODUCTION As one of the killer internet ...support of personalized and tailored learning, ...
【很好】Personalized marketing using recommendation systems_广告/传媒_人文社科...? ? ? ? 11 THE EMERGENCE OF RECOMMENDATIONS Internet-based shopping and ...
A Self-Organized Personalized Recommendation System Based on Peer-to-Peer Networks_专业资料。Abstract:Personalized recommendation systems can help people to ...
However, a theoretic idea of the ontology based personalized recommendation model for learning objects in SOA e-learning system is proposed. By providing an...
Review of personalized recommendation techniques for learners in e-learning systems Review of personalized recommendation techniques for learners in e-learning ...
to dynamically compose a fine-grained personalized recommendation for the ...Abidi, S.S.R.: Designing Adaptive Hypermedia for Internet Portals: A ...
411104 Z_X_Z2000@263.net Abstract Due to the convenience of Internet, ...In this paper,we propose a personalized recommendation system using association...
J. Kim, “A Personalized Recommendation Procedure for Internet Shopping Support,” Electronic Commerce Research and Applications, Vol.1, No.3~4, 2002. 18...
an system for personalized Internet shopping is ...The procedure for establishing this system is ... we build up a personalized recommendation system ...
Internet CDMA, CDPD, paging, Wireless Fidelity (Wi-Fi) networks Services Personalized Recommendation Well Developed Not Well Developed as ecommerce except a ...
热门文档