中文摘要：

提出一种结合用户间影响力和对象间关联关系的社会化推荐方法.该方法在建立用户兴趣模型时,利用贝叶斯方法计算用户间影响力和对象间关联关系,得到用户间影响力矩阵和对象间关联关系矩阵;然后,将其与用户-对象评分矩阵进行联合分解,解决了只考虑当前的兴趣而无法提高推荐准确率的问题.实验结果表明,所提的方法能够在推荐准确率上取得更好的效果.

英文摘要：

This paper proposed a social recommendation based on user influences and item relations on follow recommendation of weibo. The method using Bayes method to calculate influences between users and relations between items, and get the user influences matrix and item relations matrix. Then solve the problem by probability matrix factorization. Experimental results show that the proposed method outperforms other methods on recommendation accuracy.