Prediction of mortality resulted from NO 2 concentration in Tehran by Air Q+ software and artificial neural network

Abstract

In this study, for the first time, the combination of Air Q+ software and wavelet neural network was used to predict the mortality rate caused by the increase in NO 2 concentration in Tehran. In the combination of these two softwares, the wavelet neural network software was used to predict daily NO 2 concentration based on 12 effective parameters, and then the annual concentration of NO 2 was calculated using the daily concentration of wavelet neural network output. Then, annual concentration of NO 2 was used as the input of Air Q+ software. The mortality rate was calculated by Air Q+ software. In this research, the most appropriate predictive algorithm for neural network was studied and layer recurrent algorithm was the most appropriate algorithm. Then, capability of this network was enhanced to predict future NO 2 concentration by wavelet transformation, and wavelet neural network was designed. Also, NO 2 concentration is predicted for future 47 months by using of the time series of the previous data and the wavelet neural network. Analyzing the sensitivity of mortality resulted from NO 2 concentration was done by using of wavelet neural network and Air Q+ software, and it was concluded that the increase or decrease in the parameters affecting NO 2 concentration will affect the mortality rate. This research has identified petrol consumption as the most influential parameter. The conclusion is that by lowering the 10% of petrol consumption, the mortality based on NO 2 concentration in ambient air will decrease about 50%.