Abstract:

Air pollution is one of the main factors that directly threatens the quality of life. Predicting changes in air pollutants has an important role in air pollution control and management. Traditional methods such as Double Exponential Smoothing (DES) and Autoregressive Integrated Moving Average (ARIMA) used today are often insufficient for estimating air pollution. Therefore, more effective techniques are needed to determine air quality. The main purpose of this research is to develop a highly accurate air pollution prediction theory by addressing the above problems. The proposed approach is a hybrid method in which Empirical Mode Decomposition (EMD) algorithm and ARIMA model are applied together. To determine the estimation ability of the EMD-ARIMA method, data on PM10 and SO2 air pollutants in the city center of Van, Turkey, in the winter period of 2019-2020 were used. According to MAE, MAPE, RMSE and R2 performance criteria, values of 7.25 µg/m3, 20.58%, 8.84 µg/m3, 0.87 and 7.58 µg/m3, 20.73%, 8.96 µg/m3, 0.71 were obtained for PM10 and SO2 with EMD-ARIMA model, respectively. The findings reveal that the EMD-ARIMA method has a more sensitive estimation ability than traditional DES and ARIMA estimation models. The proposed hybrid approach can be used as a simple and effective method for predicting and reducing air pollution.