Experimenting user model in Cooperative IRS
Abstract
The importance of the user modeling in IRS need not longer be demonstrated. In many systems that provide access to information, the number of solutions proposed for a given query is very often extremely big. Sorting or filter these solutions are important steps for obtaining a limited relevant solutions. The main problem is how the solutions should be sorted or filtered. Moreover, sorting or filtering does not guarantee the relevance of the solutions to the user's information need. That is why it is needed to integrate the user's particularity into the methods used for calculating relevant solutions. In this paper we show how we model the user in our system METIORE and how the activities that users do can be analyzed to improve the model and the relevance of the proposed solutions