Analyzing User Behavior History for constructing user profile

User behavior history has been employed to discover userpsilas preferences and interests for personalized systems. In this paper we propose a framework for extracting user interested items by analyzing user behavior history in digital library environment. Firstly, this paper models every userpsilas behaviors as a user behavior forest (UBF). Then user behavior log is formatted with several additional attributes, and we show how to construct UBF from such logs. At last, an algorithm is presented for measuring user interests on items by considering userpsilas behavior sequence, userpsilas behavior type, and the metadata of literature.