中文摘要：

个性化推荐系统是图书馆信息系统建设的重点之一,由于读者对象不满足彼此独立的假设,阻碍了其中关联推荐计算的实施。针对上述问题,研究了基于读者活动的个性化图书馆推荐系统,给出了该系统的组成系统和处理流程,以及其关键技术;该系统将读者的动态活动数据作为个性化图书馆推荐活动的切入点,结合读者的微格式信息,最终使用新型的匹配推荐算法实现图书/读友列表的生成,从而提高推荐的高精确度。测试证明,该系统具有较高的推荐精度、覆盖度以及处理效率。

英文摘要：

Personalized library recommender systems have attracted much research attention.And analysis on reader objects don＇t satisfy the conditional independence assumption,which hinders the development of library recommendation systems.In order to deal with them,a novel recommendation system is researched with reader action processing.And its models,processing flows and key technologies are proposed as following.The system uses reader actions to generate the recommendation information,and event monitoring model to match readers and library resources.And related simulation results show that the system has better recommendation accuracy and so on than others do.