Author(s)

I. Kapageridis & A.G. Triantafyllou

Abstract

Concentrations of ambient air particles have been found to be associated with a wide range of effects on human health. PM10 concentrations are usually used as a standard measure for air pollution. Increase in the level of PM10 has been associated with increases in mortality and cardio respiratory hospitalisations. Therefore, prediction of ambient levels in certain environments is of great importance, especially in urban and industrialised areas. The present work aims to develop an adaptive system based on Artificial Neural Networks (ANN) that will allow the prediction of the maximum 24-h moving average of PM10 concentration. A special ANN architecture is employed, the Time Lagged Feed forward Network (TLFN

Keywords