Uma arquitetura para sistemas inteligentes de suporte ao usuário.

Abstract

Help desk systems for supporting clients assistance are becoming crucial for the competitive world of a globalized economy. In this context, help desk systems are computer systems playing a very mportant role in the relationship between intelligent organizations and their clients. Help desk systems are obtained by the application of Artificial Intelligence technologies such asCase-based reasoning (CBR), Knowledge-based methodologies and Data Mining. Also Internet resources, although still being in its infancy, tend to be of great value for the modelling of intelligent help desk systems. The present work deals with the combination of these resources. This dissertation has a two fold objective: firstly, we survey the current methodologies that have been used for the construction of today help desk systems: secondely. we propose a hybrid architecture for intelligent help desk systems consisting of a combination of: (i) Internet resources; (ii) CBR approach: and (Hi) Rule-based approach. In this proposed model our central effort is to promote the user interaction with the system in such a way as to equip the user himself for searching problem answers through the use of large and intensive know ledge-based systems. In the proposed architecture, solutions for problems brought by clients of organizations are stored as eitherCBR cases (the pair <problem, solution>) or heuristic rules, being the Internet the vehicle for user navigation during a questionanswering process.