融合社交和标签信息的隐语义模型推荐算法

A Recommendation Algorithm of Latent Factor Model Fused with the Social and Tag Information

为了提高隐语义模型在数据稀疏情况下推荐结果的质量,提出一种带有社交正则化项和标签正则化项的隐语义模型.根据用户社交网络和物品标签的信息,设计出描述用户和物品概况的正则化项,并利用用户对物品的历史评分计算得到用户评分偏好,将这三项引入矩阵分解目标函数中,进一步约束目标函数,最后通过梯度下降法去优化模型参数,得到推荐结果.为了验证算法的有效性,在Last.fm数据集上进行实验,实验结果表明,本文算法的推荐质量优于其他传统推荐算法.

In order to improve the recommendation performance of latent factor model under the circumstance of data sparseness, a latent factor model with the social regularization and the tag regularization is proposed.According to the user＇s social network and the item＇s tag information, the regularization depicting the profiles of the user and the item is designed, and the user rating preferences calculated by using user＇s history rating of items.These three items are introduced into the objective function of the matrix decomposition to further constrain the objective function. Finally, the gradient descent method is used to optimize the model parameters and get the recommendation result. To verify the efficacy of the proposed method, the model is tested by the Last.fm data set,and the experimental results show that the recommendation algorithm proposed in this study has a better recommendation performance compared with other traditional recommendation algorithms.