Long-term forecasting of river runoff is important for climate scientists and hydrologists. By analysing the processes of a river basin characterized by measurable variables, an empirical data-driven model can be constructed. The support vector regression technique is used in this study to analyse historical stream flow occurrences and predict stream flow values for the Amazon basin. Up to twelve month predictions are made and the coefficient of determination and root-mean-square error are used for accuracy assessment. Compared to previous studies, satisfactory results are obtained. Inclusion of environmental aspects such as precipitation and evaporation are suggested for more accurate predictions.
Reference:
Du Toit, M., Wilms, J.M., Smit, G.J.F. et al. 2016. The application of support vector regression (SVR) for stream flow prediction on the Amazon basin. SASAS Conference 2016, 31 October - 1 November 2016, Cape Town, South Africa
Du Toit, M., Wilms, J. M., Smit, G., & Brink, W. (2016). Application of support vector regression (SVR) for stream flow prediction on the Amazon basin. South African Society for Atmospheric Sciences. http://hdl.handle.net/10204/9031
Du Toit, Melise, Josefine M Wilms, GJF Smit, and W Brink. "Application of support vector regression (SVR) for stream flow prediction on the Amazon basin." (2016): http://hdl.handle.net/10204/9031
Du Toit M, Wilms JM, Smit G, Brink W, Application of support vector regression (SVR) for stream flow prediction on the Amazon basin; South African Society for Atmospheric Sciences; 2016. http://hdl.handle.net/10204/9031 .
ResearchSpace
Except where otherwise noted, this item's license is described as CC0 1.0 Universal
