Forecasting daily PM₁₀ concentrations in Seoul using various data mining techniques

Interest in PM₁₀ concentrations have increased greatly in Korea due to recent increases in air pollution levels. Therefore, we consider a forecasting model for next day PM₁₀ concentration based on the principal elements of air pollution, weather information and Beijing PM_2.5. If we can forecast the next day PM₁₀ concentration level accurately, we believe that this forecasting can be useful for policy makers and public. This paper is intended to help forecast a daily mean PM₁₀, a daily max PM₁₀ and four stages of PM₁₀ provided by the Ministry of Environment using various data mining techniques. We use seven models to forecast the daily PM₁₀, which include five regression models (linear regression, Randomforest, gradient boosting, support vector machine, neural network), and two time series models (ARIMA, ARFIMA). As a result, the linear regression model performs the best in the PM₁₀ concentration forecast and the linear regression and Randomforest model performs the best in the PM₁₀ class forecast. The results also indicate that the PM₁₀ in Seoul is influenced by Beijing PM_2.5 and air pollution from power stations in the west coast.