一种融合云模型和信任的推荐算法  

Recommendation algorithm integrating cloud model and trust

作　　者：吴旭[1] 张雄[1] 

机构地区：[1]西安邮电大学计算机学院,西安710121

出　　处：《信息技术》2016年第5期140-143,146,共5页Information Technology

摘　　要：针对传统的协同过滤推荐系统推荐精度低的问题,从用户相似度和用户信任度两个角度出发,提出了一种融合云模型和信任的推荐算法。采用云模型的相似度方法计算用户间的相似度,再采用信任推导来计算用户间的信任值,在此基础之上根据用户相似度和信任度两个因素分别确定邻居用户,得到预测评分。再将二者的预测评分结合的混合值作为最终的预测评分,从而产生推荐。实验结果表明,该算法与基于信任关系的推荐算法相比有更好的推荐效果。According to the problem of the low precision rate of traditional recommendation,from two aspects of the users＇ similarities and trusts,a recommendation algorithm integrating cloud model and trust is proposed. The cloud model similarity method is used to compute the users ＇ similarities. Trust derivation is used to compute the users＇ trusts. On this basis,users＇ similarities and trusts are used to determine the neighbor users of target user respectively,and obtain two prediction scores respectively.Then these two prediction scores are mixed to get the final prediction score. Recommendations are produced in term of the final prediction score.

关 键 词：云模型 相似性度量 信任推导 推荐系统 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]