Abstract: [Purpose/Significance] This paper targets large-scale user group of multiple services across various libraries to carry out a study on the classification and prediction of user types, aiming to provide references for the precise service construction of domestic libraries from a macro-perspective. [Design/Methodology] The questionnaire survey method was used to collect users' characteristics profile, which yielded 7,314 questionnaires from users across various libraries, including academic, public, and specialized libraries. Cluster analysis was used to classify user types based on their multi-service need characteristics. Difference tests were conducted to compare the personal, behavioral and service satisfaction characteristics of users of different types, and regression analysis was used to construct a prediction model for user types based on personal and behavioral characteristics. [Findings/Conclusion] Based on their multi-service need characteristics, this paper categorizes library users into five service need types: reading-oriented users, facility-oriented users, broad-need users, activity-oriented users, and leisure-oriented users. It is found that there are significant differences in the personal, behavioral, and service satisfaction characteristics among different types of users. We have successfully constructed a model for predicting users' service need types based on their external characteristics. [Originality/Value] This paper helps to bridge the gap in the research on the classification of user groups of multiple services across various libraries, and supports the service construction of domestic libraries in various aspects such as realizing precise services, strengthening proactive services, improving service experiences, refining service promotion, and optimizing the overall layout.

Keywords: Library service, Classification of user types, Precise service, Cluster analysis, Prediction model