4xccccc文档格式.docx
《4xccccc文档格式.docx》由会员分享,可在线阅读,更多相关《4xccccc文档格式.docx(18页珍藏版)》请在冰豆网上搜索。
10.1109/ISME.2010.157
PublicationYear:
2010,Page(s):
468-471
IEEEConferences
Abstract
|
FullText:
PDF
(255KB)
Typicaltrustfactorbasedcollaborativefilteringalgorithmisnotsuitableforuser'
smultipleinterestrecommendation.Anewalgorithmcombinestraditionaltrustfactor-basedcollaborativefilteringwithsimilaritem-basedcollaborativefilteringwaspresented.Theexperimentalresultsshowthattheproposedmethodsperformbetterthantheoldrecommendationmethod.ReadMore»
∙
ANeuralNetworks-BasedClusteringCollaborativeFilteringAlgorithminE-CommerceRecommendationSystem
JianyingMai;
YongjianFan;
YanguangShen;
WebInformationSystemsandMining,2009.WISM2009.InternationalConferenceon
10.1109/WISM.2009.129
2009,Page(s):
616-619
(276KB)
E-commercerecommendationsystemisoneofthemostimportantandthemostsuccessfulapplicationfieldofdataminingtechnology.Recommendationalgorithmisthecoreoftherecommendationsystem.Inthispaper,aneuralnetworks-basedclusteringcollaborativefilteringalgorithmine-commercerecommendationsystemisdesigned,tryingtoestablishanclassifiermodelbasedonBPneuralnetworkforthepre-classificationtoitemsandgivingrealizationofclusteringcollaborativefilteringalgorithmandBPneuralnetworkalgorithm,andcarryingontheanalysisanddiscussiontothisalgorithmfrommultipleaspects.Thisalgorithmishelpfultoimprovesparsityproblemofcollaborativefilteringalgorithmandtoformthemoreeffectiveandthemoreaccuraterecommendationresult.ReadMore»
Usingcategory-basedcollaborativefilteringintheActiveWebMuseum
Kohrs,A.;
Merialdo,B.;
MultimediaandExpo,2000.ICME2000.2000IEEEInternationalConferenceon
10.1109/ICME.2000.869613
2000,Page(s):
351-354vol.1
(488KB)
Collaborativefilteringisanimportanttechnologyforcreatinguser-adaptingWebsites.Ingeneraltheeffortsofimprovingfilteringalgorithmsandusingthepredictionsforthepresentationoffilteredobjectsaredecoupled.Therefore,commonmeasures(ormetrics)forevaluatingcollaborativefiltering(recommender)systemsfocusmainlyonthepredictionalgorithm.Itishardtorelatetheclassicmeasurementstoactualusersatisfactionbecauseofthewaytheuserinteractswiththerecommendations,determinedbytheirrepresentation,influencesthebenefitsfortheuser.Weproposeanabstractaccessparadigm,whichcanbeappliedtothedesignoffilteringsystems,andatthesametimeformalizestheaccesstofilteringresultsviamulti-corridors(basedoncontent-basedcategories).Thisleadstonewmeasureswhichbetterrelatetotheusersatisfaction.Weusethesemeasurestoevaluatetheuseofvariouskindsofmulti-corridorsforourprototypeuser-adaptingWebsite,theActiveWebMuseumReadMore»
AScalableCollaborativeFilteringBasedRecommenderSystemUsingIncrementalClustering
Chakraborty,P.S.;
AdvanceComputingConference,2009.IACC2009.IEEEInternational
10.1109/IADCC.2009.4809245
1526-1529
(2491KB)
RecommendersystemshelptoovercometheproblemofinformationoverloadontheInternetbyprovidingpersonalizedrecommendationstotheusers.Content-basedfilteringandcollaborativefilteringareusuallyappliedtopredicttheserecommendations.Amongthesetwo,Collaborativefilteringisthemostcommonapproachfordesigninge-commercerecommendersystems.TwomajorchallengesforCFbasedrecommendersystemsarescalabilityandsparsity.Inthispaperwepresentanincrementalclusteringapproachtoimprovethescalabilityofcollaborativefiltering.ReadMore»
MemeticCollaborativeFilteringBasedRecommenderSystem
Banati,H.;
Mehta,S.;
InformationTechnologyforRealWorldProblems(VCON),2010SecondVaagdeviInternationalConferenceon
10.1109/VCON.2010.28
102-107
(296KB)
WebbasedDecisionSupportsystemslikerecommendationsystemshavebecomeeffectivetoolsfordecisionmakingintherecentpast.Howevertherecommendersystemsemployingconventionalclusteringtechniques(KRS)likeK-Meansforcollaborativefiltering,sufferfromthelimitationofgettinglocaloptimumresults.ThispaperpresentsMemeticRecommenderSystem(MRS)basedonthecollaborativebehaviorofmemes.MemeticAlgorithms(MAs)areconsideredasoneofthemostsuccessfulapproachesforcombinatorialoptimization.MAsarethegeneticalgorithmswhichincorporatelocalsearchintheevolutionaryscheme.Weproposeadistinctivestrategytoperformlocalsearchinmemeticalgorithms.MRSworksin2phases-InthefirstphaseamodelisdevelopedbasedonMemeticClusteringalgorithmandinthesecondphasetrainedmodelisusedtopredictrecommendationsfortheactiveuser.RigorousexperimentswereconductedtoprovethedecisionsupportandstatisticalefficacyofMRSvisavisKRS.Resultsconfirmedthattheproposedapproachyieldsmuchbetterperformanceascomparedtotheconventionalcollaborativefilteringrecommendersystem.ReadMore»
SyncretizingContextInformationintotheCollaborativeFilteringRecommendation
RuliangXiao;
FaliangHong;
JinboXiong;
XiaojianZheng;
ZhengqiuZhang;
DatabaseTechnologyandApplications,2009FirstInternationalWorkshopon
10.1109/DBTA.2009.57
33-36
(357KB)
Socialnetworkallowsuserstoorganizecollectionsofresourcesonthewebinacollaborativefashion.Collaborativefilteringasaclassicalmethodhasbeenalsousedinhelpingpeopletodealwithinformationoverloadinfolksonomysystem.Theproblemofdevisingmethodstosolvethecontextualproblemsemergingintheprocessofrecommendationapplicationoverthesocialnetworkisincreasingopen.Hereweproposeanovelmeanstosyncretizecontextinformationintotherecommendersystem.Thispaperfirstrecalltraditionalmethodsofcollaborativefiltering,thenpresentssomedefinitionsandalgorithmframework,proposesacontextualratingestimation.Finally,experimentcomparisondemonstratesthatthecontextualapproachcanproducesbetterratingestimations.ReadMore»
AddressingInterestDiversityinP2PBasedCollaborativeSpamFiltering
FangWeidong;
DongShoubin;
GridandCooperativeComputingWorkshops,2006.GCCW'
06.FifthInternationalConferenceon
10.1109/GCCW.2006.16
2006,Page(s):
163-169
(271KB)
CollaborativeinformationfilteringtendstobeapromisingtechnologyinthefightagainInternetspam,andthepeer-to-peerframeworkisbelievedtobemoresuitabletoimplementthiscollaborationcomparedtocentralizedones.However,theassumptionofuniforminformationinterestsacrosspeersincollaborativefilteringlimitstheimprovementinfilteringaccuracyandincreasesnetworktrafficunnecessarily.Toaddressthisproblem,weproposetoconstructmultipleinterestgroupsforapeer,eachcorrespondingtoonemessagecategory;
andtouseapercolationsearchalgorithmtoretrievemessagefeedbacksinscale-freeemailnetworks.ExperimentsshowtheproposedmodelcangetremarkableimprovementinfalsepositivereductionandgoodperformanceinspamfeedbackretrievalReadMore»
LearningtheSpectrumviaCollaborativeFilteringinCognitiveRadioNetworks
HushengLi;
NewFrontiersinDynamicSpectrum,2010IEEESymposiumon
10.1109/DYSPAN.2010.5457847
1-12
(336KB)
Secondaryusersincognitiveradionetworksneedtolearnthestatisticsofspectruminordertoachieveefficientcommunications.Duetothespatialcorrelation,theefficiencyoflearningisimprovedbylettingsecondaryuserscollaborateandexchangeinformation.DuetothesimilaritybetweenthecollaborativelearningincognitiveradionetworksandtherecommendationsystemsofelectroniccommercelikeAmazon,thetechniqueofcollaborativefilteringisapplied.Predictionorientedandrewardorientedcriteriaareproposedtoderivetheprocedureofcollaborativefiltering.Fortheformercriterion,linearpredictionisusedfortheparameterestimation,heuristicmetricisderivedforchannelselection,andsimilaritybasedBoltzmandistributionisusedforcollaboratorselection.Forthelattercriterion,thetechniqueofmulti-armedbanditisappliedtomaximizethetotalrewardofspectrumaccess.Numericalsimulationshowsthattheproposedcollaborativefilteringschemecansignificantlyimprovetheperformanceofspectrumlearning.ReadMore»
AStudyontheImprovedCollaborativeFilteringAlgorithmforRecommenderSystem
HeeChoonLee;
SeokJunLee;
YoungJunChung;
SoftwareEngineeringResearch,Management&
Applications,2007.SERA2007.5thACISInternationalConferenceon
10.1109/SERA.2007.33
2007,Page(s):
297-304
(314KB)
Thepurposeofthisstudyistosuggestanal