Agente Fenix: sistema de filtragem personalizada de informações

Abstract

Nowadays Internet offers such an extensive amount of information, that it becames very difficult for the users to take advantage of it. Searching of a solution, the present work proposes a methodology of implementation of an Intelligent Agent for the personalized information filtering. The idea is to develop a set of autonomous, non-mobile, adaptative agents.The learning mechanism adopted by the agents is "relevance feedback". The search space where the infomlation will be recoveredfrom are the articles and works found in the WWW pages, selected and classified in agreement with subjects of interest. The system was developed to help the academic public, composed by teachers, graduation and masters degreé students. The initial results were quite promising. The proposed agent demonstrated to be a powerful tool of information filtering, reducing the time wasted in that activity.