Prediction of passenger volume based on Elman type recurrent neural network

Passenger volume is influenced by multiple factors and usually unable to be described with accurate mathematical models.In the paper the method of the dynamic recurrent neural network(Elman) is applied to prediction of city passenger volume.By analyzing the historic data of passenger volume of Hefei the temporal sequence of passenger volume is obtained,which can be regarded as mapping of nonlinear approximations from input to output.The Elman neural network is used to learn and train simulation experiments of the network, the output result of Elman is compared with that of the BP network,the linear regression to the output result of the network and historic data is done and the relative coefficient is obtained.The result shows that the Elman neural network is more precise and effective in passenger volume short-term prediction than the BP network.