用户多兴趣信任度的个性化推荐

目前大多数推荐技术是针对用户单方面兴趣进行的。提出了一种用户多面(multi-faced)兴趣信任度的推荐算法,以适应博客、维客、新闻文章等涉及用户多种兴趣下的推荐。新算法以一种协调的方式将传统的协同过滤算法和基于信任度的推荐算法相结合。实验结果表明,该算法不仅能适应用户多种兴趣下的推荐,而且能有效解决冷启动问题,大大提高了推荐效果。

Personalized recommendation based on trust of user multi-faced interest

Most current recommenders are aimed at the single interest, This paper provides a multi-faced interests trust recommendation algorithm so that adapts to the recommendation of users who have multi-faced interests such as bolg, wiki, news and so on. The new algorithm combines the traditional collaborative filtering algorithm and the recommendation algorithm based on the trust. The experimental result shows that the new algorithm not only fits to the recommendation of users who have multi-faced interests, but also solves the cold start problem effectively and greatly improves the recommendation effect.