An Ontology-Based Method for User Model Acquisition

Summary. This chapter illustrates a novel approach to learning user interests from the way users interact with a document management system. The approach is based on a fuzzy conceptual representation of both documents and user interests, using information contained in an ontology. User models are constructed and updated by means of an on-line evolutionary algorithm. The approach has been successfully implemented in a module of a knowledge management system which helps the system improve the relevance of responses to user queries. 1 An introduction to User Model Acquisition Research in User Model Acquisition (UMA) has received considerable attention in the last years. As the amount of information available online grows with astonishing speed, people feel overwhelmed navigating through today’s information and media landscape. The problem of information overload leads to a demand for automated methods to locate and retrieve information with respect to users ’ individual interests. Unfortunately, methods developed within information retrieval leave two main problems still open. The first problem is that most approaches assume users to have a welldefined idea of what they are looking for, which is not always the case. We solve this problem by letting fuzzy user models evolve on the basis of a rating induced by user behavior