The use of feed-forward back propagation and cascade correlation for the neural network prediction of surface water quality parameters

Modeling of surface water quality based on a deductive approach is highly non-linear, varies with time, is spatially distributed and is difficult to incorporate as part of decision-support systems. A Neural Networks (NNs) procedure provides a reliable analysis in several science and technology fields. NNs are often applied to develop statistical models for intrinsically non-linear systems. In this investigation, NNs are used in the induction of a water quality model from available field measurements for the Bahr Hadus drain in the Eastern Egyptian Delta. Two models, namely, feed-forward back propagation (BP) and cascade correlation (CC), were used. It is concluded that the CCNN model produced slightly more accurate results and learned very quickly compared with the BP procedure. The results indicated that the NNs model could be used as a non-linear dynamic system model to encapsulate site-specific knowledge and emulate the temporal sequence of one-dimensional flow systems. This NNs model undoubtedly will reduce the cost and save time in this class of problems.