Expressway traffic risk intelligent early warning method based on Bayesian network

Abstract

Expressway traffic hazards often evolve into traffic accidents. Because of the potential traffic risks and system complexity, it is difficult to deal with the real-time expressway traffic risk early warning problem by relying solely on the experience of decision makers and scattered monitoring data. Therefore, it is necessary to study the theory and method of expressway traffic risk early warning by means of data-driven decision-making approach, that is, relying on traffic big data technology to construct a holographic view of expressway traffic status for decision makers, excavating the anomalies hidden behind the data sources, and characterizing traffic accidents. This article focuses on expressway traffic risk intelligent early assessment, using the MATLAB toolbox BNT to establish a Bayesian network for expressway traffic for assessing the risk, discussing the validity and interpretability of the model. The accuracy of the training set and test set is about 0.8902 and 0.8874, respectively, which verifies the model is acceptable and valid. The innovation of this paper is to deal with the problem of expressway traffic risk early warning based on the data-driven perspective, and focuses on the interpretability of the model, giving the expressway decision makers adequate warning information.