提出一种基于《中国图书馆分类法》的用户兴趣建模方法。通过《中国图书馆分类法》分类号确定用户初始兴趣模型,使用《中图分类主题词表》中词汇描述用户兴趣主题。对于用户模型的学习,提出一种通用的方法,将各种反馈信息融合统一。在上述思想的基础上,开发出一个科技文献推荐系统,使用户在尽可能少的参与下,完成用户模型的建立、学习和更新。 This paper proposes a user interest modeling method based on “China Library Classification”. The initial interest model of the user is determined by the classification number of “Chinese Library Classification”, and the topic of user interest is described by using the words in the “Chinese Figure Classification Thesaurus”. For the study of user model, a common method is proposed, which integrates all kinds of feedback information. Based on the above ideas, a scientific and technical literature recommendation system has been developed to enable users to complete the establishment, learning and updating of user models with minimal participation.