基于动态社会行为和用户背景的协同推荐方法

Collaborative Filtering Recommendation Method Based on Dynamic Social Behavior and Users' Background Information

针对传统协同过滤推荐算法推荐精度低及冷启动的问题,提出了一种基于动态社会行为和用户背景的协同推荐方法。作为用户标注行为的结果,变化的标签体现了用户行为的动态性。该方法首先根据动态社会化标签得出用户的动态兴趣偏好相似度,然后根据用户背景信息计算出用户相似度,最后计算基于时间权重的用户评分相似度,并集成上述3个相似度找出最近邻居集,以为目标用户提供更加准确的个性化推荐。实验结果证明,该方法不仅能较好地解决数据稀疏和冷启动的问题,还能有效提高推荐算法的精确度。

To address the difficulty of data sparsity and lower recommendation precision in the traditional collaborative filtering recommendation algorithm,a new collaborative filtering recommendation method was presented based on dynamic social behavior and users’ background information.As the result of user annotation behavior,variable social tags can reflect the changes of user social behavior.Firstly,the similarities of users’ dynamic preferences are calculated based on users’ social tags.Secondly,the similarities between users are calculated based on users’ background information.Finally,the similarities of user rating are calculated based on time weight,and the above three similarities are integrated to get the nearest neighbor set for targeted users to provide more accurate individual recommendation.The experimental results show that the new method can not only improve the accuracy of recommendation,but also solve the problems of data sparsity and cold-start.