基于社交团体和用户相似度的信息推荐方法

Recommendation Method of Information Based on Social Community and User Similarity

社交网站与电子商务网站逐步实现信息共享,电子商务网站利用社交网站信息可以增强信息推荐的准确性与可信性。文章提出一种利用社交网站的用户社交网络及博文信息实现基于社交团体和用户相似度的信息推荐方法,该方法利用CNM算法发现用户所处的社会团体,通过基于语义信息的文本相似计算方法计算微博文本相似度,最后,在社团发现和文本相似度计算的基础上计算用户对项目的预测评分,实现信息推荐,并通过线下模拟实现测试该方法的有效性。

Social networking sites and e-commerce sites have gradually realized information-sharing. Using the information pro- vided by social networking sites can enhance the accuracy and credibility of recommendation system of e-commerce sites. This paper proposes a recommendation method of information based on social community and user similarity by using social networking informa- tion and micro-blog information in social networking sites. The method uses CNM algorithm to find the social community, and calcu- late the similarity of micro-blog through the text similarity calculation method based on semantic information. Finally, the paper cal- culates the prediction score of users for the recommend product based on the community detection and text similarity calculation, and tests the validity of the proposed method through the off-line simulation experiment.