A Soft Computing Fusion for River Flow Time Series Forecasting

In forecasting, the challenge of predicting river flows in time series was amongst the earliest to attract scientific interests. A broad range of mathematical approaches, from simple linear to complex non-linear methods, have been proposed in the literature for this kind of modeling. This paper introduces a hybrid method based on a soft computing fusion for river flow time series forecasting. For the experimental results reported here, this approach consistently outperformed traditional modeling methods. Findings from this specific research promise utility in the water resources and environment sector management where soft computing methods can be applied to various studies for which time series data are available.