基于用户扩展兴趣的微博推荐方法

为解决微博用户兴趣提取不准确的问题,提出一种基于用户扩展兴趣的微博推荐方法。该方法将用户个体兴趣与关联兴趣结合为用户扩展兴趣进行微博推荐。其中,用户个体兴趣从用户标签、发布微博及交互微博中提取;用户关联兴趣通过用户与其关注用户间的关注关系强度、交互频繁程度和个体兴趣相似度获取。最后,计算用户扩展兴趣与待推荐微博的相似度,对相似度降序排列产生推荐列表。实验结果表明,新方法较传统方法更具有效性和准确性。

Microblog recommendation method based on extended interest of users

In order to deal with the problem of extracting interest of microblog users inaccurately, this paper proposed a microblog recommendation method based on extended interest of users. This method combined individual interest and associated interest to represent extended interest for recommending microblogs. It extracted individual interest of users from their tags, posted microblogs and interacted microblogs. Then it got associated interest by the strength of following/followed links, interaction frequency and individual interest similarity between users and their followee. Finally, calculate the similarity between extended interest of users and the microblogs to be recommended, it generated the recommendation lists by descending the order of similarity. Experimental results show the method is more effective and precise than the traditional methods.