To accurately extract key factors from unstructured traffic accident texts and identify the interaction mechanisms among factors affecting the severity of accidents during such projects, a combined deep learning model based on BERT-BiLSTM-CRF-WApriori model is proposed. It blends Bi-directional Encoder Representation from Transformers (BERT), Bi-directional Long Short-Term Memory (BiLSTM), Conditional Random Field (CRF), and weighted Apriori association rule algorithm (WApriori),Using the BERT-BiLSTM-CRF-WApriori model, complex accident factors—including road factors, traffic factors, construction factors, environmental factors, driver factors and other complex factors in expressway reconstruction and expansion, are extracted from unstructured traffic accident texts related to expressway reconstruction and expansion projects, facilitating the analysis of interaction mechanisms influencing accident severity. The results show that the severity of sideswipe accidents in construction areas is relatively low, over speed vehicles in the lane closed construction area and involving truck will increase the severity of the accident; collisions between truck and fixture crash or intruding into the work area are likely to cause serious accidents. These research results can inform targeted measures for controlling traffic accidents and reducing injuries during expressway reconstruction and expansion project.


    Access

    Download


    Export, share and cite



    Title :

    Text Extraction and Correlation Analysis of Multi-Factor Mechanisms Influencing Traffic Accident Severity in Expressway Reconstruction and Expansion Projects


    Additional title:

    Advances in Engineering res


    Contributors:
    Chen, Gongfa (editor) / Guo, Baohua (editor) / Chen, Yan (editor) / Guo, Jingwei (editor) / Li, Jingshi (author) / Wu, Zhongguang (author) / Huang, Zechao (author) / Hao, Jiatian (author) / Zhang, Yuanbo (author) / Li, Ying (author)

    Conference:

    International Conference on Rail Transit and Transportation ; 2024 ; Jiaozuo, China October 10, 2024 - October 12, 2024



    Publication date :

    2024-12-15


    Size :

    12 pages





    Type of media :

    Article/Chapter (Book)


    Type of material :

    Electronic Resource


    Language :

    English





    Urban expressway traffic accident meteorological factor sensitivity analysis method

    WANG JIANJUN / LU XIAOJUAN / WANG SAI et al. | European Patent Office | 2021

    Free access

    Traffic accident study on Milwaukee expressway

    Lind, B.A. / Hong, H. | Engineering Index Backfile | 1965


    Analysis of Traffic Accident Characteristics of Expressway Tunnels

    Wang, Jun / Li, Feiquan / Zhao, Yifei et al. | ASCE | 2020