Combined classifier-quantifier model: A 2-phases neural model for prediction of wave overtopping at coastal structures

A 2-phases neural prediction method for wave overtopping is developed. The 'classifier' predicts whether overtopping occurs or not, i.e. q=0 or q>0. If the classifier predicts overtopping q>0, then the 'quantifier' is used to determine the mean overtopping discharge. The overtopping database set up within the EC project CLASH (De Rouck, J., Geeraerts, J., 2005. CLASH Final Report, Full Scientific and Technical Report, Ghent University, Belgium) is used to train the networks of the prediction method. The method has an overall predictive capacity, and is able to distinguish negligible from significant overtopping, avoiding large overtopping overpredictions in the area of low overtopping. The prediction method is freely available on http://awww.ugent.be/awww/coastal/verhaeghe2005. html. (C) 2008 Elsevier B.V. All rights reserved.