User modeling is a key technology of personalized recommendation system. According to the Chinese library classification, we established the user interest model, and mined the link of the information resources through the Apriori algorithm. Then we get the target user’s nearest neighbor by clustering analysis to the user through FCM algorithm and produce recommend through the collaborative filtering in the top_n nearest neighbor concentration. Experimental results show that the proposed scheme is feasible and effective, and also provide a new way for the innovation of digital library personalized service.% 用户建模是数字图书馆个性化推荐系统中的关键技术，本文依据中图法对用户兴趣进行了建模，并通过Apriori关联规则算法挖掘了信息资源各类别间的关联，然后通过FCM算法对用户进行聚类分析得到目标用户的最近邻，在Top_n最近邻居集中通过协调过滤技术产生推荐。实验结果表明，本文提出的方案是可行的和有效的，也为创新数字图书馆个性化服务提供了一种新的思路。
展开▼
