[目的/意义] 通过馆藏图书分类和流通数据，发现读者特征与馆藏流通之间的关联，建立关系模型，通过模型拟合与预测，探索读者与图书流通之间的隐含规律，为图书馆智慧管理提供技术与手段的支持。[方法/过程] 采用聚类和相关分析技术，提取读者宏观可观测特征，建立读者特征与图书分类之间直接和间接的映射关系，进而建立读者特征与分类图书流通量的回归模型，并验证模型有效性和优化模型拟合优度。根据有效模型，探索图书馆流通趋势，并挖掘读者宏观特征层面下所隐含的知识建构本质与规律，以及对图书流通产生的影响程度。[结果/结论] 具有代表读者社会角色要求的专业学习方向、代表读者间群体互动效应的入学批次、读者群体数量3个有关读者的分类特征，能够有效拟合和预测图书流通量。预测结果表明，模型准确率较高，能够作为有效工具，为图书馆开展知识服务提供可靠的技术支持。

[Purpose/significance] By means of the classification and circulation data of library collection, the paper finds the close correlation between reader characteristics and library collection circulation, establish the relationship model. And through model fitting and prediction, this study explores the implicit rule between reader and library circulation which provides technical and means support for the intelligent management of library.[Method/process] Firstly, this paper used clustering and correlation analysis techniques to extract the macroscopic observable characteristics of readers, constructed the direct and indirect mapping relationship between reader characteristics and book classification, and then constructed the regression model of the circulation of reader characteristics and classified books, and verified the validity of the model and optimized the goodness of fit of the model. According to the effective model, this paper explored the trend change of library circulation, and sum up the underlying rules of knowledge construction of the macroscopic characteristics of readers, as well as the impact on the circulation of books.[Result/conclusion] There are 3 classification characteristics of readers, namely, the professional learning direction representing the social role requirements of readers, the enrollment batch representing the interaction effect between readers and the number of readers, which can effectively fit and predict the book circulation. The prediction results show that the model has high accuracy and can be used as an effective tool to provide reliable technical support for library to develop knowledge service.