Road safety assessment method combining propensity score matching model and Bayesian model

Abstract

The invention discloses a road safety assessment method combining a propensity score matching model and a Bayesian model. The method comprises the following steps of firstly determining a test objectand pre-reference objects, and collecting corresponding parameter data of the pre-reference objects and the test object through a proper survey method; processing the pre-reference objects based on the propensity score matching model to determine an optimal reference object; and after the optimal reference object is determined, substituting the parameter data of the test object and the optimal reference object into the Bayesian model, and analyzing and judging influence effects on road safety before and after implementation of road safety management measures. According to the method, the processing is simple and easy; and through the processing of the pre-reference objects by the propensity score matching model, the reference object similar to the test object can be screened out, so that the influence of unrelated reference objects in the assessment process is reduced, the assessment accuracy is improved, and the adaptability of the Bayesian model to small sample data analysis can be enhanced.

Description

technical field [0001] The invention relates to a road safety assessment method combining a propensity score matching model and a Bayesian model. Specifically, a road section of an experiment object and a plurality of road sections of pre-reference objects are selected, and corresponding data are collected through an appropriate investigation method. The model processes the pre-reference target road section to determine the optimal reference object. Finally, based on the Bayesian method, the number of traffic accidents on the experimental road section and the optimal reference road section is compared and analyzed to judge the impact of road safety measures on road safety after implementation. Background technique [0002] With the development of social economy, the number of cars in the society is increasing year by year. At the same time, due to the complex characteristics of the road system, the number of traffic accidents is gradually increasing. How to reduce the occurre...

AI Technical Summary

Problems solved by technology

The Bayesian method fully considers the effects of confounding factors and time trend effects, and can more accurately reflect the impact of the implementation of security measures on site safety. However, the implementation of this method mainly relies on Poisson regression or negative binomial regression. Therefore, The accuracy of the assessment is greatly affected by the sample size

The propensity score matching method can ensure that the reference object and the experimental object have a similar development trend under the condition that no road safety measures are implemented, which reduces the influence of irrelevant reference objects and improves the accuracy of the evaluation. will also be restricted under the conditions

Images