Highway toll station flow prediction method based on PSO-LSSVM model

Abstract

The invention discloses a highway toll station flow prediction method based on a PSO-LSSVM model. Firstly, a toll station flow time sequence is extracted; a Pearson correlation coefficient between every two time sequences is obtained; after the flow of the toll station is ensured to have time correlation, an LSSVM model is trained through a large amount of historical data, andbased on this, the PSO algorithm is used to optimize the kernel function parameters in the LSSVM model. According to the method, the defects that a traditional linear prediction algorithm is simple in logic, too ideal inexperiment condition and large in prediction error are overcome, data are mapped to a high-dimensional space for regression through a kernel function, and the nonlinear relation between the data is considered; secondly, a least square support vector regression LSSVM model is optimized through a PSO algorithm; by selecting the optimal kernel function parameters, the prediction effect is optimal, the influence of kernel function parameter random selection or manual setting in a traditional LSSVM model on the flow prediction result of the target toll station can be eliminated through the algorithm, the stability is good, and the prediction precision is high.

Description

Technical field [0001] The invention relates to the technical field of traffic data processing and prediction, in particular to a method for predicting the flow of expressway toll stations based on a PSO-LSSVM model. Background technique [0002] As an indispensable part of modern transportation, expressway is an inevitable product of economic development and plays an extremely important role in the field of transportation. The expressway toll station is a specific entrance and exit of the closed road, so that multiple types of traffic flows are gathered here. Therefore, the flow forecast of the toll station is of great significance to traffic travelers, traffic managers, and traffic decision makers. For traffic travelers, toll station flow prediction can reasonably formulate travel routes and travel time, effectively avoid congestion and improve travel efficiency; for traffic managers, toll station flow prediction can reasonably control upstream and downstream vehicles to ensure...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the shortcomings of the traditional linear prediction method with low computational complexity, which cannot satisfy complex traffic systems and the high computational complexity of nonlinear methods, the present invention provides a traffic flow prediction method for expressway toll stations. First, the LSSVM model is used to predict , time correlation analysis is carried out on its flow, and the Pearson correlation coefficient is calculated to prove the time correlation of toll station flow; secondly, the parameters of the kernel function in the LSSVM model are optimized through the PSO algorithm, so as to obtain the optimal kernel function parameters , the algorithm can eliminate the influence of random selection or artificial setting of the kernel function parameters in the traditional LSSVM model on the flow prediction results of the target toll station, with good stability and high prediction accuracy

Images