Development of Particulate Matter Forecasting Model Using Artificial Neural Network

The pollutant emissions in ambient air led to the degradation of the health of the urban population. It is very much necessary to identify the quantum of pollution concerning the particulate matter, as it is predominant in the ambient air. This research has attempted to find the best forecasting model for the PM2.5 for the Maninagar, Ahmedabad, location using the artificial neural network, feed-forward backpropagation model. The major four gaseous criteria pollutants nitrogen dioxide (NO2), sulfur dioxide (SO2), carbon monoxide (CO), ozone (O3), and four weather parameters relative humidity (RH), wind speed (WS), wind directions (WD), and atmospheric temperature (AT) were taken into consideration while computing the forecast using ANN. 5 different combinations of pollutants and weather parameters were analyzed to get the best forecasting model for the PM2.5. This study elaborates and justifies the applicability of forecasting ‘using ANN and founds the best forecasting model which was containing all the parameters except the particulate pollutants.