Forecasting ozone pollutant (o3) in universiti pendidikan sultan idris, tanjung malim, perak, Malaysia based on monsoon using chaotic approach

Abstract

Forecasting concentration of ozone (O3) is very important because the air containing O3 can cause chronic diseases such as cancer and asthma. This study is a pilot study using chaotic approach to forecast the concentration of O3 in Malaysian education area. The studied data were the hourly O3 observed at the station located at Universiti Pendidikan Sultan Idris, Tanjung Malim, Perak. Before the forecasting model can be built, the time series are tested in advance to determine the existence of chaotic dynamics. Chaotic approach has two steps, namely the reconstruction of phase space and forecasting process. Before the phase space can be reconstructed, there are two parameters that need to be determined namely the delay time τ and embedding dimension m. Both of these parameters were obtained from the average mutual information method and Cao method, respectively. Through phase space plot and Cao method, chaotic dynamic are present in the studied O3 time series. Therefore, the forecasting model through chaotic approach using local mean approximation method is built. This chaotic approach is one of the alternative methods that can be used to forecast the O3 time series. Correlation coefficient is chosen to present the relationship between the observed value and forecasted value. The correlation coefficient for the O3 time series during Northeast Monsoon is 0.8921. Meanwhile, the value of the correlation coefficient during Southwest Monsoon is 0.9002. It is hoped that the chaotic approach can help stakeholders to manage O3 pollution in Malaysian education area.