Piecewise linear networks (PLN) for function approximation
W. Eppler, H.N. Beck · 2003
A piecewise linear network is a general neural network with three layers. It was designed for fast function approximation with a good generalization capability even in the case of very few data points. An intuitive understanding of the network processing is possible and the complexity of the network varies with the complexity of the function being approximated. This means that strong nonlinear functions are modelled by networks with more complex structure than the linear ones. The training of the network is constructive. The user provides only one parameter for the algorithm: the abort condition, when the training of the network should stop.