一种基于内容和协同过滤同构化整合推荐系统模型.pdfVIP

第36 卷 第12 期 计 算 机 科 学 V ol. 36 N o. 12 2009 年12 月 Computer Science Dec 2009 ( 西南财经大学经济信息工程学院 成都610074) 。 。 、。 R ecommender Syst em Model Based on Isomorph ic I nt egrated t o Cont ent2b ased an d Collab or ative Filterin g LI Zh on g2j un ZH OU Q2i h ai SH U A I Qin g2h on g ( School of Econ om ic In format ion E ngineering Sou th w est ern U niver sit y of F inance and Economi cs Ch engdu 6 10074 China) Abstr act Th e t wo r ecommen der sy st ems wh ich are r espect iv ely b ased on cont ent an d collab or at ive filt ering meth ods are most p opu lar . Both typ es of filt erin g meth od s h ave advant ages an d disadv ant ages . Th is p ap er pr op osed a n ew iso2 morph ic int egr at ed model an d algor ithm wh ich h ave th e merit s of th e tradit iona l r ecommender sy st ems b ased on above two meth ods and avoid th e sh or t ages of t hem t o some ext ent . Th e exp eriment al resu lt s show th at th e pr esent ed isomor2 ph ic int egrat ed model and algorithm can improv e the p erforman ce of th e tradit ion al r ecommen der sy st ems in pr edict ive accuracy . Keyword s Isomorph ic int egr at ed Cont ent2b ased Collaborat ive filt ering Recommen der syst em model [ 7] [8] [92 11] 滤推荐 ; 3) 基于知识推荐 ; 4) 基于数据挖掘推荐 ; 5) 1