基于边际价值的微博关注推荐算法

Microblog followee recommendation algorithm based on marginal value

在微博平台信息快速增长的同时,出现了信息泛滥和信息超载的问题。目前提出的关注推荐算法更多追求提高算法的准确性,而忽视了推荐系统的多样性,使用户获得的推荐列表过于相似,信息冗余度高。为提高推荐系统的多样性,基于传统协同过滤推荐和LDA的文本挖掘方法,本文提出了一种基于边际价值的微博关注推荐算法,以新增关注用户的边际价值作为推荐标准,期望用最少的关注用户满足最大的信息需求。实验结果表明,当推荐用户数为10时,本文提出算法的多样性达到了64.72%,而传统的协同过滤算法只有38.96%,推荐列表的多样性有明显提高。

While the information on the microblog platform is growing rapidly,there is a problem of information overload. At present,more recommendation algorithms focused on improving the accuracy of the algorithm,while ignoring the diversity of the recommendation system,so that the recommendation list obtained by the user is too similar and the information redundancy is high. In order to improve the diversity of recommendation system,based on traditional collaborative filtering recommendation and text mining method based on LDA topic model,this paper proposes a microblog followee recommendation algorithm based on marginal value,with the newly added user’s marginal value as the recommendation standard and lowering the similarity of the recommendation list. Experiments show that when the number of recommended users is 10,the diversity of the algorithm proposed in this paper reaches 64.72%,while the traditional collaborative filtering algorithm is only 38.96%,and the diversity is significantly improved.