Author(s)

P. Mlakar and M. Boznar

Abstract

Air pollution forecasting models are a helpful tool for controlling pollution around sources such as large thermal power plants. Recently we developed a neural network based short-term SC^ pollution forecasting model for the Sostanj Thermal Power Plant. One of the most important problems that should be solved in order to improve the model performances is the determination of features. We developed several methods for feature determination. The methods are suitable for both neural network based models and for classical statistical models as well, so they can be widely used in the field of air pollution forecasting. Beside this mostly heuristic technique we adapted three techniques known from pattern recognition theory for the use in the field of air pollution modelling. These techniques are saliency m

Keywords