Forecasting Transport Mode Sharing in Beijing

摘要

Unreasonable traffic structure is proven to be a great contributor to environmental problems in Beijing. In order to make suggestions for traffic structure optimization, a two-step numerical research combining the regression analysis and the BP neural network has been made in this study. The first step is to establish the regression analysis model for the transport mode shares and, more importantly, to identify the relevant factors influencing the share rates of transport modes. The other step is to build a forecasting model based on a BP neural network, taking the identified relevant variables into account. The data of the involved variables are collected from Beijing Transport Annual Reports. We use the data from 2000 to 2011 to build the models, and use the data of 2012 to validate the models. The prediction results are shown to be accurately consistent with the real-gained data, and therefore the models are validated.