Traffic flow prediction based on generalized neural network

This work presents an intelligent neuron model, which is based on linearly independent functions and sigmoid function with adjustable parameters. It is proved that the information storage ability of this intelligent neuron is greatly improved compared with traditional ones, consequently greatly improves the information processing ability of the whole neural network. Meanwhile, this paper forms a generalized neural network model by these intelligent neurons, and uses this generalized neural network to predict traffic flow data of DaLian city. Experiment shows that the results predicted by this generalized neural network are greatly superior to the ones predicted by traditional back-propagation neural network, and meet the practical requirements well.