Predicting Carbon Monoxide Concentrations in the Air of Pardis City, Iran, Using an Artificial Neural Network

摘要

We applied the MLP neural network and the RBF neural network using air quality data from Pardis City to predict CO concentrations. The key points of the results are as follows: 1. The MLP neural network with two hidden layers reached a suitable result in predicting CO concentrations. The first and second hidden layers in the MLP neural network contained 13 and 25 neurons, respectively. Also, the best error values achieved using this network of MAE, VE, RMSE, and MBE in the RBF neural network were 0.08, 6.11, 0.095, and 0.0611, respectively. 2. The coefficient of determination, IA, and E between the observed data and the predicted data using the MLP were 0.96, 0.9057, and 0.957, respectively, which indicates the MLP neural network's accuracy in predicting CO concentrations in Pardis City.