《Applying memetic algorithmbased clustering to recommender system with high sparsity problem》.pdfVIP

J．Cent．South Uni、乙(2014)21：354l一3550 垒springer DoI：10．1007／sll771—014．2334-4 memetic to Applyingalgorithm-basedclustering recommenderwith systemhighsparsity problem MARI烈G Sansanee2 ukritl，THEERA一啪0NNiponl，AUEPHANwIRIYAKUL ofElectrical of Mai Mai 1．D印a舳ent 50200，Tllailand； Engiflee豳g，FacukyEngifleering，ChiangUniverSi吼Chiang of of Mai Mai 2．Depamnem 50200，Thailalld ComputerEngineering，FacuhyEng协ee咖g，ChianguniVers时，Chiang ◎Cen仃alSoum Pressand Berlill 14 UniVersity Springer-VerlagHeidelbe唱20 Abstract：Anewrecornmendationmemodwas basedonmemetic memodwas presented algorittlIll．basedclusteriIlg．Thepmposed onf011r tested real．worlddatasets．Itsrecommenda“on isevaluatedand withthatofthe hi曲lysparse perfbmlance compared intemlsof methods 行equency-based，user．based，item-based，肛meansclustering-based，锄dgeneticaIgorithm-based precision， Fl score．TheresultSshowmatme memodbetter undermenewusercold．start recall，and proposed yields perfbmlance Dmblem wheneachofnewactiveusersselectsoneortwoitemsinto出e F follr only