结合似然关系模型和用户等级的协同过滤推荐算法.PDFVIP

计算机研究与发展 ISSN N 111777TP J ournal of Comp ut er Research and Develop ment 45 (9) : 1463 1469 , 2008 结合似然关系模型和用户等级的协同过滤推荐算法 高 　滢 　齐 　红 　刘 　杰 　刘大有 (吉林大学计算机科学与技术学院 　长春 　1300 12) (吉林大学符号计算与知识工程教育部重点实验室 　长春 　1300 12) (gying @j lu . edu . cn) A Collaborative Filtering Recommendation Algorithm Combining Probabil istic Relational Model s and User Grade Gao Ying , Qi Hon g , Liu J ie , and Liu Dayo u ( Col leg e of Comp ute r S cience an d Technology , J i l i n Uni ve rs i ty , Chang ch un 1300 12) ( M i nis t ry of E d ucation , Key L aboratory of S y m bol ic Comp ut ation an d K now le d g e Eng i nee ri ng , J i l i n Uni vers i ty , Chang ch un 1300 12) Abstract 　Collaborative filt erin g i s one of succe ssf ul t echnolo gie s for buil ding reco mmen der sy st em s. U nfort unat ely , it suff er s f ro m sp ar sit y and scalabilit y p roblem s. To addre ss t he se p ro blem s , a collaborative filt erin g reco mmen dation algorit hm co mbinin g p robabili stic relational mo del s an d u ser grade ( PRMU GCF) i s p re sent ed . PRMU GCF ha s p rimary t wo p art s. Fir st , a u ser grade f unction i s defined , an d u ser grade ba sed collaborative filt erin g met ho d i s u sed , w hich can find neighbor s for t he t ar get u ser only in hi s near grade , and t he number of can didat e neighbor s can be co nt rolled by a p aramet er , so reco mmendatio n efficiency i s increa sed and it solve s t he scalabilit y p ro blem . Second , in or der to u se variou s kin ds of infor matio n for reco mmendation , u ser grade ba sed collaborative filt ering met ho d i s co mbined wit h p robabili stic relational mo del s ( PRM ) , t h