Use of neural networks in the identification of pollutant concentration scenarios in the city of Rio de Janeiro

Abstract

Studies carried out in several countries have reported an association between air pollution and several indicators of morbidity and mortality, even when pollutant concentrations are below standard limits. This work has as geographic place of study the city of Rio de Janeiro,and aims to identify local and global scenarios characterized by high or low pollution days and typical or atypical weather days. It also verifies the associations with the statistical distribution of health event counts. The method used to identify these scenarios was Kohonen topological maps based on neural networks.

Keywords

• Poluição do ar. Mapas de Kohonen. Redes Neurais.