A User Modeling Server for Determination of Semantically-Enriched User Interests

This paper describes a user modeling server capable of harvesting  user interests in a nonintrusive manner based on the content accessed by users through mobile devices. The server enables mobile agents to log information about the content accessed by the user either in form of URL, text markup, or URI of an ontology concept referring to an encountered concept. Based on the information about occurrences of terms and semantic relations among them, the server clusters the terms into three groups, namely, interested, partially interested, and not interested.