中文摘要：

针对用户的兴趣变化具有时间敏感性特点,文章提出基于用户兴趣变化的数字图书馆知识推荐模型。首先融合标签和时间等因素,通过用户使用标签的频率以及对资源的标注时间等信息构建用户-资源评分矩阵;然后结合协同过滤算法,计算目标用户最近邻从而完成知识推荐,并在此基础上设计个性化知识推荐服务模型;最后探讨系统知识推荐服务机制及其应用。

英文摘要：

In view of the time sensitivity of users＇ interest drift, this paper presents a knowledge recommendation model for digital library based on users＇ interest drift. Firstly, it colligates tags and time factors, through the usage frequency and marking time of tags to construct a user resource evaluation matrix, then combines with collaborative filtering algorithm, and calculates the target user＇s nearest neighbor set and conducts knowledge recommendation. On the basis of the above considerations, it designs the personalized knowledge recommendation service model. Finally, knowledge recommendation mechanism and application of system are discussed.