Adaptive-network-based fuzzy inference system (ANFIS) modelbased prediction of the surface ozone concentration

This paper presents the results of the tropospheric ozone concentration modeling as the dependence on volatile organic compounds - VOCs (Benzene, Toluene, m,p-Xylene, o-Xylene, Ethylbenzene); nonorganic compounds - NOx (NO, NO2, NOx, CO, H2S, SO2 and PM10) in the ambient air in parallel with the meteorological parameters: temperature, solar radiation, relative humidity, wind speed and direction. Modeling is based on measured results obtained during the year 2009. The measurements were performed at the measuring station located within an agricultural area, in vicinity of city of Zrenjanin (Serbian Banat, Serbia). Statistical analysis of obtained data, based on bivariate correlation analysis indicated that accurate modeling cannot be performed using linear statistics approach. Also, considering that almost all input variables have wide range of relative change (ratio of variance compared to range), nonlinear statistic analysis method based on only one rule describing the behavior of input variable, most certainly wouldn’t present accurate enough results. From that reason, modeling approach was based on Adaptive-Network-Based Fuzzy Inference System (ANFIS). Model obtained using ANFIS methodology resulted with high accuracy, with prediction potential of above 80%, considering that obtained determination coefficient for the final model was R2=0.802.