ABSTRACT
7.1 Introduction
An artificial neural network (ANN) is a computing paradigm tailored to mimic natural neural networks (Haykin 1999). It can be defined as "a computational mechanism able to acquire, represent, and compute a mapping from one multivariate space of information to another, given a set of data representing that mapping" (Garrett 1994). A typical ANN comprises an input layer that receives inputs from the environment, an output layer that generates the network's response, and some intermediate hidden layers. Maier and Dandy (2000) furnished a comprehensive review on using neural network models to predict and forecast water resources variables. The basis of ANNs is our current understanding of the brain and its pertinent nervous systems. Numerous processing elements connected by links of variable weights are grouped together to constitute black box representations of systems. In this chapter, the characteristics of ANNs and the commonly used backpropagation forward-feeding ANN are delineated. Two real applications of ANN are also demonstrated. The first application case study presents the analysis of algal dynamics data from a coastal monitoring station. The second application is for prediction of long-term flow discharges in Manwan based on historical records.
