Performance Comparison of Haze Prediction Using Chaos Theory and Multiple Linear Regression

Abstract

Forecasting haze is essential for protecting the environment, the economy, and public health. It assists authorities in taking preventative action to lessen the adverse effects of haze episodes and boost community resistance to air pollution. The goal of this study was to create a model for haze prediction by using two methods, multiple linear regression and chaos theory. In this study, chaos theory forecasts haze using univariate time series which is PM10, whereas multiple linear regression (MLR) utilizes multivariate time series for its predictions, namely ambient temperature, wind speed, ozone, nitrogen dioxide, carbon monoxide, and sulphur dioxide. Data for this study will be collected during the southwest monsoon from an industrial area in Klang, Selangor. The results of these two models will be compared to determine which model gave better performance. With these predictive models, policymakers and relevant authorities can receive timely alerts, allowing them to implement preventive measures that can reduce the impact of haze on public health and the environment.