New cooperative approach for dynamic classification of multimodal and evolutionary data

Abstract

In this paper we present a new approach for dynamic classification of multimodal and evolutionary data, based on a multi-agent architecture. This approach is adaptive with the evolutions of classes and data in order to optimize the allocation of entities from disparate sources into clusters, and to strengthen the mechanism of incremental classification so that determinate a cases of creating new cluster. This approach will allow the classifiers agents to collaborate in making final decisions. It was implemented on the platform JADE, where every step is handled by specialized agents working together and communicating between them.