A Prediction Model of Short-term Passenger Flow for Urban Transit Hubs

Accurate prediction of the passenger flow plays a very important role in preparing advanced organization schemes and contingency plans for urban transit hubs .Therefore ,a combinational prediction model based on BP neural network and Least Squares Support Vector Machine (LSSVM ) is proposed in this paper .First ,a BP neural network is a-dopted to present an initial prediction based on the historical passenger volume .Then ,the LSSVM model is used to refine the "initial prediction"to reach the final predicted passenger volumes at urban transit hubs .The experiment results of this paper show that the proposed model can improve the prediction accuracy of the passenger flows at urban transit hubs by 1% ,which shows that the model in this paper can overcome the uncertainty caused by a single model .