Dynamically updating UIDB to implement personalized service

User interests in electronic commerce reflect the behavior set of users acting on certain impulse. Electronic commerce web stations (ECWS) might make full use of intelligent IT to create and refine user interests database (UIDB) to make services personalized. In this paper, the issues about how to implement such kind of services are investigated, the concepts of user interests and their transitions are defined; the structure of UIDB and how to create it are explored, and the analysis and mining of the data in server log files to help dynamically updating UIDB are discussed in detail. Then, ECWS can actively recommend suitable series of pre‐sent web pages for different users and flexibly deal with transitions of users' interests. The dynamic structure may make the system perfect after a period of use to help ECWS to offer their users with personalized service.