A User Interest Model for Web Page Navigation
M. TAMER ÖZSU · 2003
Making recommendation requires predicting what is of interest to a user at a specific time. Even the same user may have different desires at different times. It is important to extract the aggregate interest of a user from his or her navigational path through the site in a session. This paper concentrates on the discovery and modelling of the user’s aggregate interest in a session. This approach relies on the premise that the visiting time of a page is an indicator of the user’s interest in that page. The proportion of times spent in a set of pages requested by the user within a single session forms the aggregate interest of that user in that session. We first partition user sessions into clusters such that only sessions which represent similar aggregate interest of users are placed in the same cluster. We employ a model-based clustering approach and partition user sessions according to similar amount of time in similar pages. In particular, we cluster sessions by learning a mixture of Poisson models using Expectation Maximization algorithm. The resulting clusters are then used to recommend pages to a user that are most likely contain the information which is of interest to that user at that time. Although the approach does not use the sequential patterns of transactions, experimental evaluation shows that the approach is quite effective in capturing a Web user’s access pattern. The model has an advantage over previous proposals in terms of speed and memory usage.