Reasonable prediction of trip mode choice of urban residents is the foundation of effective implementation of traffic demand management and traffic control strategy. Combined with trip survey data of residents, this paper first analyzes main factors influencing trip mode choice of urban residents. Then, prediction models are built respectively based on a multinomial logit model and probabilistic neural network. By using these two models, probability of each trip mode choice is calculated and residents' final choice is estimated. Finally, an accuracy test of these two models is completed by calculating their hit rates and errors. The results indicate that the prediction model built based on probabilistic neural network can enhance prediction accuracy and overcome the inherent defects of a multinomial logit model. Therefore, it can predict trip mode choice of urban residents more effectively.
Research on Prediction of Traffic Mode Choice of Urban Residents
11th International Conference of Chinese Transportation Professionals (ICCTP) ; 2011 ; Nanjing, China
ICCTP 2011 ; 449-460
2011-07-26
Conference paper
Electronic Resource
English