Adaptive Agent-Based Architecture for Health Data Integration

Abstract

The work presented in this paper is developed in the context of the “PersoDiagMedi” project which is Franco-Tunisian cooperation between laboratories’ multidisciplinary programs in the fields of computer science and health-care. In the health-care domain, Artificial Intelligence (AI) provides multiple technologies that allow machines to learn, act and make decisions autonomously. In this sense, AI helps experts and doctors in diseases’ diagnosis and the detection of emerging diseases’ presence. However, the medical data come from multiple sources: doctors, biologists, meteorological specialists, and environmental organizations. The difficulty of the epidemic state’s surveillance lies in the conciliation between the search for the largest number of relevant signals and their treatments. In this paper, we propose the project’s general architecture that facilitates medical data integration and data processing to detect unusual facts and to prevent the presence of emergent diseases. In this proposed data integration architecture, we present the different functionalities, describe its layers and components as well as present an adaptive multi-agent system for the unusual facts’ detection. To this end, the feasibility of our proposal is first proven and later two use cases are presented to cover users’ needs.