Efficient Management of User Interests for Personalized Communication Services

Nowadays, there is a lot of interest in context- aware services, especially those services taking into account user interests. An important example is the automated filtering of the overwhelming amount of information available to the user. There is a clear need for the automated capturing of user interests without explicit user interaction. In this paper we present a set of algorithms for the management of user interests. This information is used for updating a tree structure, with added weight values, representing user interests. To illustrate the approach taken, we detail the use case of an instant messaging communication service which updates the user profiles and searches for content matching the topic of the ongoing conversation and the specific interests of the user. The retrieved content is recommended to the user, who is then able to provide feedback by selecting the content he prefers. The gain of the detailed approach is illustrated by means of simulation results, taking real-time constraints into account.