Prediction of Future Ozone Concentration for Next Three Days Using Linear Regression and Nonlinear Regression Models

• Main Authors: Abdullah M.M.A.-B (Author), Ezzah Shamimi, R. (Author), Hafiz Zawawi, M. (Author), Mohamad, D. (Author), Mubin Zahari, N. (Author), Rahim S.Z.A (Author), Rosli M.F (Author), Saad M.N.M (Author), Talib M.S (Author), Zain A.M (Author), Zia Ul-Saufie, A. (Author)
• Format: Article
• Language: English
• Published: Institute of Physics Publishing 2019
• Subjects: Air pollution; Coefficient of determination; Errors; Forecasting; Green computing; Index of agreements; Linear regression; Linear regression models; Mean square error; Nonlinear regression models; Normalized absolute errors; Ozone; Performance indicators; Prediction accuracy; Root mean square errors
• Online Access: View Fulltext in Publisher; View in Scopus
• ISBN: 17578981 (ISSN)
• DOI: 10.1088/1757-899X/551/1/012006

The aim of this research is to predict the ozone concentration level for the next three days. Linear regression model and a nonlinear regression model are used to measure the air pollution data and the result was compared. The performance indicator used to evaluate the accuracy of the methods is Index of Agreement (IA), Prediction Accuracy (PA) and Coefficient of Determination (R2). While Normalized Absolute Error (NAE) and Root Mean Square Error (RMSE) are for error measured. The results show that the prediction for the next three days. The highest ozone concentration of the linear regression model is 0.085ppm at Petaling Jaya, Selangor. While the lowest concentration for the linear regression model is 0.015 ppm at Klang, Selangor. Besides, the highest ozone concentration for the nonlinear regression model is 0.1 ppm at Petaling Jaya, Selangor for the second-day prediction. Comparison between the linear regression model and a nonlinear regression model indicates that nonlinear regression model can as an alternative method to the linear regression model. © 2019 Published under licence by IOP Publishing Ltd.