Research On Simulation Prediction Method Of Traffic Mode Sharing Based On Probit Model

The probit model is one of the models of the disaggregate prediction method for the traffic mode split.Because it involves high-dimensional integral operation,it is very difficult to directly solve the model.In response to this problem,some scholars have proposed a Monte Carlo simulation algorithm for the probit model.The algorithm has the problem of utility fixed term parameter calibration and utility random term generation,which is difficult to apply the probit model in practical work.Based on the above situation,this research studies the simulation and prediction method system of traffic mode sharing based on the probit model,and uses modern computing technology,statistical analysis software and Python programming tools to implement the method.Firstly,this research analyzes the principle of Monte Carlo simulation method of Probit model and the functions of mathematical statistical software,analyzes the feasibility of method implementation from the level of technical means,and proposes the corresponding methodological ideas and technical framework.On this basis,the utility fixed term parameter calibration,the utility random term generation method,and the set counting processing method of the non-set counting method are studied respectively.A parameter calibration method with the process of traffic mode influence factor selection,model parameter calibration,and utility function establishment,and a utility random term generation method using Box-Muller inverse conversion method,as well as a Monte Carlo simulation set counting processing method based on Probit model are established.Finally,the Monte Carlo simulation algorithm is programmatically implemented using Python.To investigate the prediction effectiveness of the simulation method,this study conducted an example validation using travel survey data of City S.The results show that the model parameters are well calibrated and the overall traffic mode choice prediction hit rate is 90.7%.The calculated percentages of 37.4%,36.1% and 26.5%for private car traffic,public transportation and non-motorized trips in the city,respectively,were obtained with 90% accuracy compared to the actual trips,which verified the effectiveness of the method of this study.