Adaptive user profiling for deviating user interests

Abstract

Although the World Wide Web has evolved to be an enormous source of information over the last decade, finding the required information poses a significant challenge to users as there is simply too much information available. Therefore systems that enable users to perform personalised information searches are becoming increasingly important. This paper presents a mechanism which provides recommendations based on a user profile that adapts to the changing interests of the user. Taking the multiple and diverse user interests into consideration, such as the short-term, long-term and periodic interests, the proposed mechanism constructs a user profile that evolves according to these interests. Articles from different news sources are presented to the users that they have subscribed to through an RSS aggregator and are ranked according to the user's profile. The mechanism adapts the user profile according to the changing user interests automatically and makes recommendations based on this profile.