A Comparative Analysis of Artificial Neural Network and Support Vector Regression for River Suspended Sediment Load Prediction
The artificial neural network (ANN) model and support vector regression (SVR) model have gained tremendous popularity among the researchers during the past couple of decades. Both of the models are very powerful in prediction and have several applications in different fields, which also include suspended sediment load prediction. In this work, the predictive capability of ANN and SVR model is investigated to estimate the daily suspended sediment load (SSL) in Tawang Chu River, Jang of Arunachal Pradesh, India. The performance of the models is evaluated using three quality measuring parameters, i.e., mean squared error (MSE), root-mean-square error (RMSE), and mean absolute error (MAE). From the experimental results, one can conclude that the predictive capability of SVR is better compared to ANN in terms of all of the quality measuring parameters.
KeywordsArtificial neural network Support vector regression Suspended sediment load Prediction
This work was fully supported by the Science and Engineering Research Board, Government of India (SERB), under early career research award ECR/2016/001464. We are also thankful to NHPC LIMITED, Tawang Basin Project to provide this data.
- 7.Botsis, D., Latinopulos, P., Diamantaras, K.: Rainfall-runoff modeling using support vector regression and artificial neural networks. In: 12th International Conference on Environmental Science and Technology (CEST2011) on 2011, pp. 8–10Google Scholar
- 12.Khan, M.Y., Hasan, F, Tian, F.: Estimation of suspended sediment load using three neural network algorithms in Ramganga River catchment of Ganga Basin, India. Sustainable Water Resources Management, pp. 1–7 (2018)Google Scholar
- 15.Rumelhart, D.E., Hinton, G.E., Williams, R.J.: Learning internal representations by error propagation. California Univ. San Diego La Jolla Inst. for Cognitive Science on 1985. La JollaGoogle Scholar
- 16.Vapnik, V., Mukherjee, S.: Support vector method for multivariate density estimation. In: Advances in Neural Information Processing Systems on 2000, pp. 659–665. MIT Press, CambridgeGoogle Scholar