Abstract: This research expands the methods and research perspectives of user access behavior
analysis, and provides support for the services and management decision of libraries. This research uses
the 2017-2018 Fudan University Teachers and Students Chinese Electronic Journal Resource Access
Behavior Dataset, constructs user access behavior profiles by data mining techniques and methods,
and explores and discovers users regularities in access time, operating preferences, access patterns, and
retrieval preferences. It draws the conclusion that the active time of users shows a cyclical change of “three
increases and three decreases” within a day, and changes periodically with work, study and vacation
within one year; in terms of operating preferences, users prefer instant and quick short visits; a visit is
often performed alternately between simple retrieval and browsing operations, and users are increasingly
inclined to obtain information through online browsing instead of downloading; in terms of retrieval
methods, users prefer to use the default “subject” as the limiting condition, and directly use the retrieval
box on homepage to perform simple retrieval.