用户与资源信息发掘深度不足阻碍了推荐服务在高校图书馆中的应用。针对上述问题,提出了基于多源信息融合的高校图书馆推荐系统的基本原理与结构、组成模块与运行流程,并给出了基于灰色关联度的信息融合与功率谱估计的推荐关键技术与方法。实验结果证明,该系统具有较高的读者覆盖率、准确度以及召回率,为图书馆的网络化与自动化建设提供了新思路。 The lack of deep exploration of users and resources hinders the application of recommendation service in college library. In view of the above problems, this paper proposes the basic principle and structure of university library recommendation system based on multi-source information fusion, the module and operation flow, and presents the recommended key technologies and methods of information fusion and power spectrum estimation based on gray relational degree . The experimental results show that the system has high readership coverage, accuracy and recall rate, which provides a new idea for the network and automation construction of the library.