摘要
随着社会化媒体的兴起,信息资源的数量呈现爆炸式增长,如何在海量的信息中帮助用户发现有用的知识成为亟需解决的问题.社会化推荐方法作为一种有效的信息过滤技术,由于能够结合社会网络的特点,模拟现实社会中的推荐过程,在分析用户历史行为的基础上,主动向用户推荐满足他们兴趣和需求的信息,受到了研究者们的广泛关注.但目前已有的方法大都只从用户间社会关系的角度出发,仅认为相互信任的朋友间具有相似的兴趣爱好,而忽略了推荐对象间的关联关系对推荐结果产生的影响.针对以上存在的问题,文中从推荐对象间关联关系的角度出发,假设具有关联关系的推荐对象更容易受到同一用户的关注,并进而在已有的社会化推荐算法的基础上,提出了一种结合推荐对象间关联关系进行推荐的算法.算法使用共享的潜在特征空间对目标函数的求解过程进行约束,使其在考虑用户间社会关系的同时,也考虑到推荐对象间关联关系所起到的重要作用.实验结果表明,与主流的推荐算法相比,文中所提出的方法在分类准确率和评分误差等多种评价指标上都取得了更好的结果.
With the advent of social media and the exponential growth of information generated by online users,how to help users find useful knowledge from vast amounts of data has become the major problem to be solved.Social recommendation method as one of the effective information filtering techniques attempting to provide active suggestions with social networks has been well studied.Most of these methods assume trusted friends have similar interests.Typically,they simulate the recommendation process in real social networks to automatically predict the user＇s preference by collecting the history behaviors and ratings from his/her friends.However,these methods only consider the influence of social networks from users＇ perspective,and assume items are independent and identically distributed.This assumption ignores the fact that item relations can be important factors in many recommendation scenarios.Aiming at solving the above problem,based on the intuition that related items will be probably selected by the same user,we propose a novel social recommendation method and incorporate item relations using a probabilistic matrix factorization framework from the items＇ perspective.Specifically,our method utilizes the shared latent feature space to constrain the objective function,and considers the influence of user connections and item relations simultaneously.Experimental results in real world social network show that the proposed approach outperforms state-of-the-art recommendation algorithms in terms of precision and rating error.
出处
《计算机学报》
EI
CSCD
北大核心
2014年第1期219-228,共10页
Chinese Journal of Computers
基金
国家自然科学基金(61272240
60970047
61103151)
教育部博士点基金(20110131110028)
山东省自然科学基金(ZR2012FM037)
山东省优秀中青年科学家科研奖励基金(BS2012DX017)资助~~
关键词
社会网络
矩阵分解
推荐系统
协同过滤
社会化推荐
social networks
matrix factorization
recommender system
collaborative filtering
social recommendation
