Non-invasive Browser Based User Modeling Towards Semantically Enhanced Personlization of the Open Web

Currently the user´s web search is disjoint from the resources which is subsequently browsed. Specifically the related instances of the search are not displayed on the following pages. This lack of continuity between the actual search and the web sites displayed may lead to skimming by the user to identify what is relevant on the pages. This paper presents an approach to the continuous modeling of a user´s interests through a browser based plug-in that may be used to inform the intelligent highlighting of content on the browsed pages. Specifically the paper presents two related systems, one for capturing and modeling the user´s browsing behavior and the second for leveraging the power of linked data to highlight information relevant to the user´s interests. The design, implementation and evaluation of both systems are presented as well as a conclusion and outlook for further work.