结合矩阵分解的混合型社会化推荐算法  

Hybrid socialized recommendation algorithm based on matrix factorization

作　　者：杨丰瑞[1,2] 刘彪 杜托[1] Yang Fengrui;Liu Biao;Du Tuo(Communication Technology Applications Resarch Center,Chongqing University of Posts ＆ Telecommunications,Chongqing 400065 China;Chongqing Information Technology（Group）Co.Ltd,Chongqing 401121,China)

机构地区：[1]重庆邮电大学通信新技术应用研究中心,重庆400065 [2]重庆重邮信科集团股份有限公司,重庆401121

出　　处：《计算机应用研究》2018年第6期1631-1635,共5页Application Research of Computers

基　　金：重庆市研究生科研创新基金资助项目(CY15166)

摘　　要：推荐系统是用来解决当今时代信息过载的重要工具。随着在线社交网络的出现和普及,一些基于网络推荐算法研究的出现已经引起研究者的广泛关注。然而大多数信任感知的推荐系统忽略了用户有不同行为偏好在不同的兴趣域。考虑用户间特定域信任网络,并且结合推荐项目之间特征属性信息,提出了一种新型社会化推荐算法(H-PMF)。实验表明,H-PMF算法在评分误差和推荐精度上都取得了很好的效果。Recommender systems（ RSs） have become important tools for solving the problem of information overload. With the emergence and popularity of online social networks,some studies on network-based recommendation algorithm have emerged,raising the concern of many researchers. Trust is one kind of important information available in social networks and is often used for performance improvement in social-network-based RSs. However,most trust-aware RSs ignore the fact that the user has different preference in different domains of interest. This paper proposed a new social recommendation algorithm（ H-PMF）,which not only considered the user-specific domain trust network,but also combined the feature attribute information between recommended items. Experiments show that the H-PMF algorithm has better performance in both scoring error and recommendation accuracy.

关 键 词：信任网络 协同过滤 矩阵分解 推荐系统 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]