Nonlinear Autoregressive (NAR) Forecasting Model for Potomac River Stage using Least Squares Support Vector Machines (LS-SVM)
Nian Zhang1, Tilaye Alemayehu2, Pradeep Behera3
1Dr. Nian Zhang, Department of Electrical and Computer Engineering, University of the District of Columbia, 4200 Connecticut Ave. NW, Washington, DC, 20008, USA.
2Mr. Tilaye Alemayehu, Department of Electrical and Computer Engineering, University of the District of Columbia, 4200 Connecticut Ave. NW, Washington, DC, 20008, USA.
3Dr. Pradeep Behera, Department of Civil Engineering, University of the District of Columbia, 4200 Connecticut Ave. NW, Washington, DC, 20008, USA.
Manuscript received on 30 January 2015 | Revised Manuscript received on 12 February 2015 | Manuscript Published on 28 February 2015 | PP: 47-51 | Volume-4 Issue-9, February 2015 | Retrieval Number: I1979024915/15©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: This paper investigates the ability of a least-squares support vector machine (LS-SVM) model to improve the accuracy of streamflow forecasting. Cross-validation and grid-search methods are used to automatically determine the LS-SVM parameters in the forecasting process. To assess the effectiveness of this model, streamflow records from Geological Survey (USGS) gaging station 1652500 on Four Mile Run of the Potomac River, were used as case studies. The performance of the LS-SVM model is compared with the recurrent neural networks model trained by Levenberg-Marquardt backpropagation algorithm. The results of the comparison indicate that the LS-SVM model is a useful tool and a promising new method for streamflow forecasting. Index Terms—
Keywords: Water Quantity Prediction, Least Squares Support Vector Machine
Scope of the Article: Support Vector Machine