An Application of Artificial Neural Networks for Air Quality Prediction of SO2 Concentrations in Calabar Metropolis

Ekechukwu Christopher Chinasa; Joseph Abebe Obu; Eucharia C. Okoro

Abstract

The past few years have seen intense enhancements in our aptitude to simulate complicated physical systems using computers, which has led to the use of artificial neural networks (ANNs) in various capacities of scientific research. The artificial neural network has been applied to various ecological pollution complications and which has revealed some degree of precision. The study aims to advance a neural network air quality forecast model for the complex district of Calabar Metropolis within Cross River State, Nigeria. In this study, two forecast models are established using a feed-forward neural network for air pollutant SO2. Metrological data such as rainfall, humidity, temperature, wind direction and wind speed are given as input parameters while the concentration of SO2 was considered as the output variable in this study. The enactment of the developed model was assessed through a measure of Mean Square Error (MSE) from the fabricated networks architecture, the best forecast performance was observed in a model with network structure 05–10–01 and MSE 0.00245

Abstract

Keywords