Akbar Rahimi
Mir sattar Sadrmousavi
Masoud Geravanchizadeh
Reza Nabavi

Forecasting of air quality parameters is one topic of air quality research today due to the health effects caused by airborne pollutants in urban areas. This work deals specifically with the use of an artificial neural network (ANN) for hourly NO2 and NOx modeling in the Tabriz metropolitan. The development of an ANN model is presented to predict the NO2 and NOx concentrations as a function of meteorological conditions. The optimum structure of ANN was determined by a trial and error method. It was found that the structure of ANN with 30 neurons in the hidden layer had the best performance. Log sigmoid transfer function was used in hidden layer and pure line transfer function was used in output layers. The correlation coefficient (R2) was 92% of the variability in the NO2 and 94% of the variability in the NOx concentrations. In this paper, the main emphasis is a studying the hourly NO2 and NOx concentrations during October 2003.

Air pollution prediction; Artificial neural network; Multilayer perceptron; Tabriz metropolitan; Multiple regression model; NO2; NOx