Logistics forecasting using improved fuzzy neural networks system

In this paper, we proposed and trained a fuzzy neural network system to estimate future logistics demand. The structure of neural network in the system is similar to that of BP network, except that here the nonlinear sigmoid functions in the networks are replaced by fuzzy reasoning process and wavelet functions respectively. Moreover, the trained network system is put into practical logistics demand forecasting. The experimental results show that it has good properties such as a fast convergence, high precision and strong function approximation ability and is good at predicting future logistics amount.