ABSTRACT
In numerous disciplines related to applied science and technology, an increasing trend is observed towards the use of artificial neural networks (NN) for non-linear modelling, control, optimisation, design, data analysis and classification. For recent applications in areas such as meteorology, oceanography, hydraulics, hydrology and ecology the reader is referred to: Wüst (1995), Minns (1996), Minns and Hall (1996), Scardi (1996), Abrahart and Kneale (1997), Recknagel et al. (1997), Clair and Ehrman (1998), Hsieh and Tang (1998), Lange (1998), Sanchez et al. (1998), Shen et al. (1998), Van Gent and Van den Boogaard (1998), Wen and Lee (1998) or See and Openshaw (1999). NN are in most cases used to perform an identification of input-output relations and multi layer perceptrons (MLP) or radial basis function networks (RBFN) are the two most popular types of tool for studies; see Haykin (1994), Beale and Jackson (1990) or Abrahart (Chapter 2).
