Author(s): Mojtaba Shourian; Ahmad Khazaie Poul

Linked Author(s):

Keywords: No Keywords

Abstract: Reliable and precise predictions of river flows are of major concern in water resources analysis like reservoirs'designing. In the current study, common data driven models including multi linear regression (MLR) as a statistical method, artificial neural network (ANN) as non-linear ones and eventually K-nearest neighbors (KNN) as a non-parametric and easy learning one, are applied to forecast monthly flows in St. Clair river between US and Canada. In the developed methods, three scenarios are defined in order to study the effect of different input vectors. Performance of models is evaluated using MAE, RMSE, and R parameters as statistical indices. Also results of these scenarios confirmed both ANN and KNN as good models among applied methods for long term river flow forecasting.

DOI:

Year: 2016