Quantitative Research Of Expert Opinions And Application In Mega Event Passenger Flow Forecast

Since the reform and opening-up,an increasing number of mega events have started to land in large and medium-sized cities,along with the rapid development of domestic economy,continuously improving international status and increasingly perfect urban functions.Prediction of mega event passenger flow quantum and typical day passenger flow volume is the significant component of traffic demand forecasting as well as the important basis of transportation planning.Precise prediction of its scale is in favor of the successful hosting of activity and the value realization.Through comparing expert consultation method with other methods and expounding the application of this method in the passenger flow prediction of International Horticultural Exposition in Beijing,it is found that it could use expert knowledge and experience to define a reasonable range with flexible,handy,efficient and time-saving advantages.Meanwhile,the method is equipped with such defects as non-convergent prediction conclusion and difficult quantification.In this connection,depthly investigation into expert opinion quantification and its application into mega event passenger flow prediction.Firstly,this paper teases domestic and foreign literature,elaborating the traffic characteristics of mega events.It defines the research scope,summarizing the research status and deficiencies of forecasting methodologies.In the next place,through expounding the application of prediction method in the passenger flow of International Horticultural Exposition in Beijing,it reveals that utilizing expert consultation and review to define passenger flow scale is a qualitative analysis process;expert opinions are short of quantitative analysis.Then,through analyzing and comparing the frequently-used quantitative methods of expert opinions,it points out that latent dirichlet allocation can improve the defects of expert opinion analysis,quantify and analyze expert opinions and establish the quantitative analysis model of expert opinion.At last,it uses expert opinion training model,calibration parameter,and training model to carry out the quantitative analysis of expert opinions and extract latent topics.It defines the relative importance of transportation planning facilities,stating the application of expert opinion quantification analysis in the passenger flow of International Horticultural Exposition in Beijing.The main contents of this paper are as follows.(1)Research on the quantitative analysis of expert opinions: qualitative and quantitative methods are adopted to predict passenger flow scale.The rationality of passenger flow scale is defined through expert consultation and review.Through analyzing and comparing the frequently-used expert opinion quantification methods,it reveals that strong subjectivity,poor convergence of prediction results and lack of quantitative analysis exist in expert opinion analysis.In order to overcome these defects,the analysis model of expert opinion is set up.(2)Combination scene of road carrying capacity feedback: Chinese words segmentation,stop-word removal and other semantic analysis theories are adopted to conduct structured processing of expert opinions;the processed texts are imported into the latent dirichlet allocation program;Gibbs sampling algorithm is adopted to extract the topic-feature word distribution.Combining the relative importance of topic and feature word probability distribution computing planning facilities,it can provide feedback about the main scenes in the road bearing capacity combination scenario.(3)Rolling prediction model of group expert decision-making: on the basis of passenger flow prediction process,when the qualitative or quantitative prediction conditions change,the predicted value of the method changes accordingly.It shows that it needs the new round of expert consultation to further define the passenger flow prediction model.Through the expert opinion quantification analysis of the feedback combination scene of road carrying capacity under new conditions,it promotes the prediction model to be close to the actual passenger flow need,forming the rolling prediction model of group expert decision-making.