Simulating Hourly Air Quality Index Using the Classification and Regression Tree (CART)

Air quality index (AQI) is a proxy of air pollution which is useful to recognize the effects of air pollutants on human health in urban areas. The AQI is a generic information tool assists public health authorities to manage their monitoring over any diseases causing by air pollution. Consequently, the main objective of the present study is to simulate hourly AQI through Classification and Regression Tree algorithm (CART) using air pollutants and meteorological variables. The estimation of AQI through USA Environmental Protection Agency (USEPA) method using hourly observed concentration of air pollutants. The measured variables were based upon hourly slots throughout 2007 and 2008 at the Gholhak monitoring station. A regression tree model was then assigned to simulate hourly AQI. Several statistical descriptors such as RMSE, MAE, MSE, FB, IA, R and R2 used to evaluate the performance of the regression tree model. Results showed that the regression tree model has a good ability (IA=1) to simulate the hourly AQI. The results proved that the CART method can be applied successfully as a tool for decision making and problem solving for better air quality management. Finally air quality modeling and estimation would be helpful to concerned authorities in providing the necessary information to the general public, to protect their health and take necessary precautionary measures.

Keywords:

Air pollution , Air quality , CART algorithm , Simulation , Tehran