Ensemble of neural predictors for forecasting the atmospheric pollution

Author

Siwek, Krzysztof ; Osowski, Stanislw ; Garanty, K. ; Sowinski, M.

Author_Institution

Warsaw Univ. of Technol., Warsaw

fYear

2008

fDate

1-8 June 2008

Firstpage

643

Lastpage

648

Abstract

The paper presents the application of an ensemble of neural predictors for forecasting the daily meteorological PM₁₀ pollution. The support vector machine has been used as the basic predicting network. The bagging technique has been applied to adapt different predictors. The results of many predictors have been combined together to form final forecasting. The blind source separation has been applied as the integration tool. The results of forecasting of the real pollution measured in the northern region of Poland have been presented and discussed.

Keywords

air pollution; blind source separation; forecasting theory; meteorology; neural nets; support vector machines; atmospheric pollution forecasting; bagging technique; blind source separation; meteorology; neural predicting network ensemble; support vector machine; Air pollution; Atmospheric measurements; Bagging; Blind source separation; Cities and towns; Meteorology; Pollution measurement; Support vector machines; Weather forecasting; Wind forecasting;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 2008. IJCNN 2008. (IEEE World Congress on Computational Intelligence). IEEE International Joint Conference on

Conference_Location

Hong Kong

ISSN

1098-7576

Print_ISBN

978-1-4244-1820-6

Electronic_ISBN

1098-7576

Type

conf

DOI

10.1109/IJCNN.2008.4633862

Filename

4633862