Application of combined neural networks in nonlinear function approximation

Discusses neural networks in combination with mathematical models used to approximate functions whose forms are unknown. The method can speed up the convergence and improve the generalization ability of neural networks. The validity of the method presented is proved by two practical examples.