Comparative study of forecasting techniques for Dengue cases

Abstract

Dengue is a viral infection that spreads rapidly and is endemic in more than 100 tropical and subtropical countries in Africa, America and Asia-Pacific regions. Generally, the epidemiology of dengue is influenced by a complex interplay of factors including rapid urbanization and increased population density, capacity of health systems, effectiveness of vector control systems, urban cleanliness, etc. In Brazil, which is considered a tropical country, there is an increasing incidence of cases in recent years, specifically in the capital Pernambucana Recife there is a favorable environment for the significant increase in dengue cases as presented by the Epidemiological Bulletin of the Health Surveillance Secretariat of the Ministry of Health. The present work presents a comparative study of Data Mining techniques following the CRISP-DM approach, for a model for predicting Dengue cases in Recife-PE.