Authors: Yara S. Tadano, Thiago Antonini Alves, Natalia S.S. Silva, Hugo S. Valadares

Air pollution has been investigated worldwide, due to its health risks. The health risk analysis due to air pollutants emissions and climate changes are commonly performed using statistical regressions. However, an innovative alternative consists on using Artificial Neural Networks (ANNs). Thus, this work aims to compare the performance of different ANNs on the analysis of air pollution health impacts. As case study, it was analyzed the impact of particulate matter with aerodynamic diameter less or equal to 10 μm (PM10) and climate changes (temperature and humidity) on the number of hospital admissions by respiratory diseases in Campinas city/Brazil. The experimental data of atmospheric pollution and meteorology were obtained from air quality monitoring stations located in the study area. The ANNs, Extreme Learning Machines and Echo State Networks was applied successfully.

Yara S. Tadano | Thiago Antonini Alves | Natalia S.S. Silva | Hugo S. Valadares
Universidade Tecnológica Federal do Paraná, Ponta Grossa, PR, Brasil