Probability Models Describing Hazard Identifi...

Probability Models Describing Hazard Identification Effectiveness of Workers in the Construction and Maintenance Industry

Author Name : Shishir Dewangan, Manish Kumar Mishra, Jaykant Gupta

ABSTRACT

This study examines the hazard identification effectiveness of workers in the construction and maintenance industry through the application of various probability models. The dataset used provides a comprehensive overview of accidents, including the count of accidents in each country, location, industry sector, accident level, potential accident level, gender category, employee/contractor status, and critical risk category. To assess hazard identification effectiveness, four probability models, namely Binomial Distribution, Poisson Distribution, Negative Binomial Distribution, and Neural Network, are employed. Each model generates insights into the probability of successful hazard identification and, in some cases, the probability of failed hazard identification. The results reveal the performance of different models, providing valuable information for decision-making regarding hazard identification strategies in the construction and maintenance industry. The findings demonstrate that the Binomial Distribution model yields a probability of successful hazard identification of 0.43, while the Poisson Distribution model produces a higher probability of 0.89. In contrast, the Negative Binomial Distribution model indicates a lower probability of 0.10. However, it is important to note that the Neural Network model presents a probability of successful hazard identification of 0.67, but does not provide a value for the probability of failed hazard identification. These outcomes contribute to a better understanding of hazard identification effectiveness and offer insights into the strengths and limitations of different probability models. The results can inform decision-making processes aimed at enhancing hazard identification practices and improving overall worker safety in the construction and maintenance industry.

Keywords: hazard identification, probability models, Binomial Distribution, Poisson Distribution, Negative Binomial Distribution, Neural Network, successful hazard identification, failed hazard identification