Abstract
Air quality monitoring and forecasting tools are necessary for the purpose of taking precautionary measures against air pollution, such as reducing the effect of a predicted air pollution peak on the surrounding population and ecosystem. In this study a single Feed-forward Artificial Neural Network (FANN) is shown to be able to predict the Air Pollution Index (API) with a Mean Squared Error (MSE) and coefficient determination, R2, of 0.1856 and 0.7950 respectively. However, due to the non-robust nature of single FANN, a selective combination of Multiple Neural Networks (MNN) is introduced using backward elimination and a forward selection method. The results show that both selective combination methods can improve the robustness and performance of the API prediction with the MSE and R2 of 0.1614 and 0.8210 respectively. This clearly shows that it is possible to reduce the number of networks combined in MNN for API prediction, without losses of any information in terms of the performance of the final API prediction model.
Highlights
Air quality is monitored continuously and manually to detect any changes in the ambient air quality status that may cause harm to human health or the environment
The single forward Artificial Neural Network (FANN) network with a single hidden layer was applied with the Levenberg-Marquardt training algorithm with a sigmoid activation function in the hidden layer and a linear activation function in the output layer
The structure of the single FANN is represented by the number of nodes in each layer
Summary
Air quality is monitored continuously and manually to detect any changes in the ambient air quality status that may cause harm to human health or the environment. Several methodologies for the assessment and monitoring of air pollutants have been implemented by organizations such as the Department of Environment (DOE) of Malaysia which has developed indexes for air quality In response to this concern, several studies on air quality prediction using artificial neural networks have been done [2, 3]. Gardner and Dorling [13] and Perez and Trier [14] have adopted this model to predict the NO and NO2 concentration based on meteorological data in Central London and traffic junctions in Santiago City in Chile respectively They have concluded that the MLP has better performance compared to their previously developed regression models.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
