一种基于遗忘规律的微博用户关注度计算方法

Micro-blog user attention degree method based on forgetting law

准确把握微博用户所关心的领域对于目标营销和个性化推荐具有重要意义。利用用户历史关注信息预测用户将来可能关注的领域,是一种解决该问题的有效思路。此过程中,若用户关注某一领域的信息越多说明用户将来关注该领域的可能性越大;关注某信息的时间距离当前时间越近,则该信息的参考价值越大。这两点与人类对知识重复学习强化记忆和逐渐遗忘的过程非常相似,因此以人类遗忘规律为基础,提出了一种微博用户关注度计算方法。真实数据集上的实验表明,该方法能够很好地预测用户关注度分布情况。

It is significant for targeted marketing and personalized recommendation to grasp accurately the interest of micro- blog user. It is a valid idea to predict current attention by history information of the user. In this process, the higher attention de- gree to the field is, the more a user pay attention to a concern field, and the stronger the reference value of the message is, the nearer the distance between user attention time to a message and current time is. These two points can be regarded as the process of human＇s repeated learning and forgetting gradually. Therefore, this paper proposed a method of calculation micro-blog user＇s attention based on forgetting law. Experiments show that the method can predict micro-blog user＇s attention better.