Transport Modal Split of Commercial Sites Based on Artificial Neural Network

On the basis of the characteristics of the transport mode structure at commercial sites, this paper initially analyzes the factors that have influenced public transport modal split. Taking advantage of the strong nonlinear mapping and generalization characteristics of Back-Propagation (BP) neural network, a prediction model of public transport trip proportion is established based on investigation data of commercial sites in Beijing. Meanwhile, this paper makes quantitative analysis of the factors and forecasts public transport trip proportion at commercial sites under different policies.