Taking into account Tabbed Browsing in Predictive Web Usage Mining

摘要

Over the last few years, browser tabs have become a very common tool for web users and have been extensively used to perform parallel navigations. Tabbing facilitates web browsing but results in an imbrication of navigations, which makes it more difficult to understand users' behavior. That is why very recent research has been focusing in analyzing this new kind of usage. This work follows a previous publication in which a new model was proposed to model parallel browsing. In this paper, we propose a new strategy to better take into account tabbing activity. Experiments are performed on an open browsing dataset. Results show that our model provides an accuracy similar to the one of a state-of-the-art model that implicitly takes into account parallel browsing. It thus constitutes a strong basis to estimate tabbing activity. We then present the statistics about parallel browsing that our approach provides.