Daily Prediction of PM10 using Radial Basis Function and Generalized Regression Neural Network

Due to urbanization and population growths air pollutant (AP) are increasing drastically. Therefore, predictive AP models have become an important tool to provide air quality management for the site. In this study prediction of PM10 is performed using radial basis function neural network (RBFNN) and generalized regression neural network (GRNN). For this daily average value of PM2.5, NO, Benzene, vertical wind speed and PM10 of Ardali bazar in Varanasi, India are considered. RBFNN and GRNN incorporates input variables as PM2.5, NO, Benzene, vertical wind speed and target variable as PM10. It is found that RBFNN predict better than GRNN and multiple linear regression models.