Citation

Deng XJ. Adv. Transp. Stud. 2023; (SI 2): 53-64.

Copyright

(Copyright © 2023, Arcane Publishers)

DOI

unavailable

PMID

unavailable

Abstract

The traditional traffic accident feature modeling methods cannot eliminate the influence of feature dimensionality in the feature extraction process, resulting in low prediction accuracy. This article proposes a feature modeling method for traffic accidents on foggy highways based on multi-source data fusion. Extract spatial features of traffic accidents based on highway traffic flow images, and refine the constructed feature vector set. Mining the correlation between various feature factors and establishing correlation rules as the initial parameters for feature modeling. Adopting multi-source data fusion algorithms to achieve feature fusion and constructing a traffic accident feature fusion model. The experimental results indicate that the model can be used for high-precision prediction of the number of traffic accidents.

Language: en

Keywords

Crash Factors; Crashes; Models; Road Safety