Different Approaches of Multiple Linear Regression (MLR) Model in Predicting Ozone (O3) Concentration in Industrial Area

Abstract

Meteorological conditions and other gaseous pollutants generally impacted the development of ozone (O₃) in the atmosphere. The purpose of this study was to create the best O₃ model for forecasting O₃ concentrations in the industrial area and to determine the variables that affect O₃ concentrations. Five-year data of meteorological and gaseous pollutants were used to analyze and develop the prediction model. Based on three distinct techniques, three separate multiple linear regression (MLR) prediction models of O₃ concentration were developed. MLR₃ had the highest correlation coefficient of 0.792 during development as compared to models MLR₁ and MLR₂. MLR₂ was deemed the best O₃ prediction model, however, since it had the lowest error values of root mean square error (3.976) and mean absolute error (3.548) when compared to other models. The establishment of an O₃ prediction model can offer local governments with early information that could help them reduce and manage air pollution emissions.