Mining Web Logs of Online News Service for User Profiling
User profiling plays an important role in online news recommendation systems. In particular, the quality of news recommendation services depends on the user model. In this paper, we analyze the relationship between users clicking behaviors and the category of the news story to model users interests by mining web log data of an adaptive news systems of Idolcan. We train a Memory-based User Profile (MUP), which imitates human beings learning, remembering and forgetting mechanisms, to predict users potential interests dynamically. We mainly focus on experimental analysis to refine the MUP scheme. Firstly, we materialize the meanings of all parameters of MUP by important factors (i.e., absorbing factor, forgetting factor, timescale and learning strength) in human beings learning and forgetting process. Secondly, we demonstrate how to determine the values of parameters for different users to reflect their distinct learning and forgetting abilities. Thirdly, we derive a threshold from MUPs recursion formula, which can be used to simply distinguish long-term and short-term interests. Our evaluations are carried out on Idolcans web log data, results show MUP can model users profile effectively.
User profile Memory model Learning and forgetting curve
Wei Wang Dongyan Zhao
Institute of Computer Science & Technology, Peking University, Beijing, China Department of Electron Institute of Computer Science & Technology, Peking University, Beijing, China
国际会议
2011 International Conference on Database and Data Mining(ICDDM 2011)(2011年数据库和数据挖掘国际会议)
三亚
英文
65-69
2011-03-25（万方平台首次上网日期，不代表论文的发表时间）