Vessel traffic flow forecasting based on BP neural network and residual analysis

Through the study on the vessel scheduling problem of the port, it is found that the navigational environment is complicated near the channel area. In order to provide better services, the traffic flow information should be clearly managed. This paper builds a vessel traffic flow forecasting model based on Back Propagation (BP) neural network, and validates the model with some Automatic Identification System (AIS) data from September 20, Tianjin Port, and then uses the residual analysis to test the reliability of the model and the applicability of the sample data. Firstly, the initial data should be normalized, and next, determine the basic parameters of the BP neural network prediction model; Secondly, BP neural network simulation is carried out with SPSS Modeler 18.0 software, and the residual value was analyzed by SPSS20.0 software; Finally, after analyzing residuals, simulation results, comparing the predicted values with actual values, the results indicates that the forecasting model can accurately predict the traffic flow of the port and it also can provide the basis need for the port berth scheduling and channel resource management.