• Login
    View Item 
    •   DSpace Home
    • ASCE
    • Journal of Management in Engineering
    • View Item
    •   DSpace Home
    • ASCE
    • Journal of Management in Engineering
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Role of National Conditions in Occupational Fatal Accidents in the Construction Industry Using Interpretable Machine Learning Approach

    Thumbnail
    View/Open
    JMENEA.MEENG-5516.pdf (2.944Mb)
    Date
    8/4/2023 1
    2023-08-04
    Author
    Kerim Koc
    Metadata
    Show full item record
    Abstract
    Current national occupational safety and health (OSH) initiatives follow reactive approaches, i.e., if it breaks, fix it. Existing accounts, however, failed to improve national OSH performances substantially, which imposes the need for an in-depth and proactive (fix it so it will not break) investigation of national occupational fatality risks. Despite many studies examining the fatality risk of workers based on project-, company-, and/or behavior-related factors, the role of national conditions on the countrywide fatality risk of workers has not been explored. The present research leverages the national statistics of Turkey to examine their influence on construction workers’ fatality risk through a machine learning–based prediction model. Several widely used machine learning methods were adopted for determining whether the upcoming month poses a significant fatality risk for construction workers or not based on national statistics of the previous month. According to analysis results, the gradient boosting decision tree algorithm yielded the highest prediction performance in terms of f1-score. The recently developed game theory–based Shapley Additive Explanations (SHAP) algorithm was used to identify whether and how national conditions affect countrywide fatality risk of construction workers. Findings illustrate that the share of the construction sector in employment, market demand, and labor shortage are the most significant national factors in determining the fatality risk. SHAP summary and SHAP dependence plots are further presented to provide decision makers with a clearer understanding of hidden relationships between fatality risk and national conditions. In addition, a framework that can be practically used by policy makers and governmental authorities is developed to help minimize national occupational fatality risk. Overall, predicting national fatality risk in the industry and identifying the national precursors of occupational fatalities contribute to the development of macrolevel safety improvements based on country-specific conditions.
    URI
    http://yetl.yabesh.ir/yetl1/handle/yetl/4293982
    Collections
    • Journal of Management in Engineering

    DSpace software copyright © 2002-2015  DuraSpace
    Contact Us | Send Feedback
    Theme by 
    @mire NV
     

     

    Browse

    All of DSpaceCommunities & CollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

    My Account

    LoginRegister

    Statistics

    View Usage Statistics

    DSpace software copyright © 2002-2015  DuraSpace
    Contact Us | Send Feedback
    Theme by 
    @mire NV