社交网络环境下基于信任的推荐算法

现有的基于信任的推荐算法通常假设用户是单一和同质的，没有充分挖掘信任关系信息，且相似关系和信任关系的融合缺乏高效的模型，极大地影响了推荐的准确性和可靠性。本文提出一种基于信任的推荐算法。首先结合全局信任和局部信任，并利用信任的传播性质对信任关系进行建模，然后设置推荐权重，综合考虑相似度和信任度来构建用户间的偏好关系，筛选出邻居。接着将基于记忆的协同过滤思想和社交网络的信任关系融入概率矩阵分解模型，同时使用自适应权重动态决定各部分的影响程度，形成高效统一的可信推荐模型Trust-PMF。本文的算法在FilmTrust、Epinions这两个数据集上与相关算法做了对比验证，结果证实了此算法的高效性。

Trust-Based Recommendation Algorithm in Social Network

The existed trust-based recommendation algorithms usually assume that users are homogeneous,which haven''t fully mined the trust relationship information.Moreover,the lack of efficient model for the integration of similar relationship and trust relationship greatly affects the accuracy and reliability of the proposed model.Therefore,this paper proposes a trust-based recommendation algorithm called Trust-PMF.It combines similarity with trust to build user''preference and select the target user''s neighbors.Then the probability matrix factorization model is extended by integrating memory-based idea and trust information,meanwhile,a dynamic adaptive weight is used to determine the degree of influence of each part so as to form a unified and efficient Trust-PMF model.Finally,the experiment results on FilmTrust and Epinions data sets demonstrate that our method outperforms the state-of-the-art methods.