Using Consensus Methods to User Classification in Interactive Systems
Janusz F. Sobecki, Ngoc Thanh Nguyên · Advances in intelligent and soft computing · 2002
In this paper we describe the consensus problem solution applied to the web-based system interface construction for each class of users. We assume that we have a multiagent system that collects knowledge about user interactions with different systems in the form of user profiles which serve to user classification. This multiagent system also collects the interface profiles of all users of these systems, which are used for the adaptive interface construction. Because of the differences among the users and their experiences, the knowledge stored in the form of interface profiles is inconsistent, even for the users belonging to the same class, so the most efficient interface could be found when we reconcile the knowledge by means of consensus methods.