基于认知度的协同过滤推荐.pdfVIP

ISSN 11-1777／TP 计算机研究与发展 1000—1239／CN and 5—519，2009 Journalof ResearchDevelopment Computer 46(Suppl．)：51 基于认知度的协同过滤推荐算法 李建国 姚良超 汤 庸 郭 欢 (中山大学计算机科学系广州510275) (1jguo@mail2．sysu．edu．cn) Collaborative Recommendation CognitionDegree—Based Filtering Algorithm Li GuoHuan Jianguo，YaoLiangchao，TangYong，and Yat—sen 510275) Science，Sun (DepartrnentofComputer University，Guangzhou is usedin AbstractCollaborative widely recommender filtering many make onunrateditemsfortheactiveuser hisindividualon predictions bycomparing ratings algorithms rateditemswiththeotherusers’．Inthis novel is paper，a memory—basedalgorithmproposed，in areclassifiedtheir usersareselectedand are whichtheitems attributes，similar by key ratings fortheactiveuserafter resultsonMovieLens predicted computingcognitiondegree．Experimental datasetshowthatthis theothertraditional collaborative algorithmoutperforms memory—based filteringalgorithms． wordscollaborative Key filtering；keyattribute；cognitiondegree；memory—basedalgorithm； MovieLens 摘要 协同过滤技术被广泛应用于各种推荐系统当中．基于内存的协同过滤算法通过比较目标用户与 其他用户的已有评分，为目标用户的未评分项目作出相应的预测．提出了一种新的基于内存的算法。根 据项目的关键属性对它们进行分类，通过计算用户对各类项目的认知度，为目标用户选择相似用户并预 测评分．通过Moviel