Network mining techniques to analyze the risk of the occupational accident via bayesian network

• Published in: International journal of system assurance engineering and management Vol. 13; no. Suppl 1; pp. 633 - 641
• Main Authors: Nayak, Nihar Ranjan, Kumar, Sumit, Gupta, Deepak, Suri, Ashish, Naved, Mohd, Soni, Mukesh
• Format: Journal Article
• Language: English
• Published: New Delhi Springer India 01-03-2022; Springer Nature B.V
• Subjects: Accident data; Accident prediction; Bayesian analysis; Construction companies; Construction industry; Contingency tables; Engineering; Engineering Economics; Frequency analysis; Logistics; Machine learning; Marketing; Occupational accidents; Occupational safety; Organization; Original Article; Quality Control; Reliability; Risk analysis; Safety and Risk; Tabulation; Univariate frequency analysis; Potential severity; Occupational accident; Cyclical probability; Cross-tabulation; Network mining; Machine learning; Bayesian network; Spreadsheet application
• ISSN: 0975-6809; 0976-4348
• DOI: 10.1007/s13198-021-01574-1

Today, as the construction industry grows, the frequency of occupational accidents has risen as well. The advancement of technology, inadequacies in workplace safety procedures, and untrained workers are the primary causes of these workplace mishaps. In this research, occupational accident data were preprocessed and then subjected to univariate frequency and cross-tabulation analysis. As a consequence of the research, risk factors for occupational accidents were identified. Then, using Bayesian networks, the impacts of these factors on occupational accidents were examined (BNs). A Bayesian network is a graphical model that captures the conditional dependencies between variables. On a dataset from an international construction firm, the Bayesian network was deployed. Finally, we evaluated the correctness of the constructed Bayesian network and other performance criteria, as well as the model's efficacy. The experimental findings indicate that utilizing machine learning methods, some occupational accident situations may be predicted with great accuracy. The main aim of the paper is to aims to get rid of the repetitive patterns in the data and present a more reasonable level of data for the classification analysis.