MODELING METABOLIC ENERGY BY NEURAL NETWORKS

The apparent metabolic energy (EMA) of barley is modelled as a functio n of 12 easily obtainable analytical parameters by applying neural net works with the error back-propagation learning strategy. Kohonen maps and Ward's clustering technique have been used to define the objects f or the training and test sets. The architecture of the neural network and the relevant parameters of error back-propagation learning have be en optimised providing a RMS of 1.081 and a correlation coefficient (p redicted versus found values) of 0.82. Contour maps of all variables i ncluding the output EMA value have been obtained by applying the count er-propagation learning strategy in a two-layer neural network. The re sponses yielded by the networks show that this method is capable of es tablishing a quantitative relationship between EMA and the original va riables.