Personalisation of information based on the concept of relevance using a user model
Résumé
In information retrieval systems (IRS), the user's interests and preferences are obtained in most cases implicitely through the user's interaction with the system. In our studies, we represent each user by a user model, which is composed of four parameters : the user's identity, his behaviour, the history of his past activities and his feedback. For each application of an IRS, we request the user to provide explicitely his main objective and provide reasons for accepting or refusing a solution given by the system in order to obtain a knowledge of the user's interests and preferences. Ous proposals are implemented in a prototype METIORE for accessing publications of our research laboratory.