Support Vector Regression Algorithms in the Forecasting of Daily Maximums of Tropospheric Ozone Concentration in Madrid
In this paper we present the application of a support vector regression algorithm to a real problem of maximum daily tropospheric ozone forecast. The support vector regression approach proposed is hybridized with an heuristic for optimal selection of hyper-parameters. The prediction of maximum daily ozone is carried out in all the station of the air quality monitoring network of Madrid. In the paper we analyze how the ozone prediction depends on meteorological variables such as solar radiation and temperature, and also we perform a comparison against the results obtained using a multi-layer perceptron neural network in the same prediction problem.
KeywordsSupport Vector Machine Support Vector Regression Meteorological Variable Multilayer Perceptron Tropospheric Ozone
Unable to display preview. Download preview PDF.
- 3.Garfias-Vázquez, M., Audry-Sánchez, J., Garfias-Ayala, F.J.: Tropospheric ozone prediction in Mexico city. Journal of the Mexican Chemistry Society 49(1), 2–9 (2005)Google Scholar
- 13.Smola, A.J., Schlkopf, B.: A tutorial on support vector regression. Statistics and Computing (1998)Google Scholar