Neural predictor ensemble for accurate forecasting of PM¹⁰ pollution

The paper presents the neural network approach to the accurate forecasting of the daily average concentration of PM₁₀. Few neural predictors are applied: the multilayer perceptron, radial basis function, Elman network and support vector machine. They are used for prediction either in direct application or in combination with wavelet decomposition, forming many individual prediction results that will be combined in an ensemble. The important role in presented approach fulfills the integration of this ensemble. We have proposed solution applying the principal component analysis and additional neural network responsible for final forecast. The numerical experiments for prediction of the daily concentration of the PM₁₀ pollution in Warsaw are presented. They have shown good overall accuracy of prediction in terms of all investigated measures of quality: the errors RMSE, MAE MAPE as well as index of agreement and correlation of the prediction and true values.