Prediction of Human Error in Maritime Operations Machine Learning Approaches to Fatigue, Workload, and Complexity Analysis of Navigation

Resumen

This study aims to analyze the factors that affect Human_Error in maritime operations using logistic regression. The main factors studied include Fatigue, Workload, Experience, and Navigation Complexity, as well as the interaction between these variables on the possibility of human error. The study draws on human reliability theory, which highlights that human error is influenced by fatigue, workload, and the complexity of the work environment. The method used was binary logistic regression with 50 observations, where Human_Error being the dependent variable. The analysis was carried out in two stages: (1) testing the direct effects of Fatigue and Workload and (2) testing the moderation effects of Experience and Navigation Complexity. The results showed that Fatigue and Workload increased the probability of Human_Error, with odds ratios of 1.848 and 1.721, respectively. However, the moderation model was insignificant, with a concordant percentage of only 59.3%, indicating a less accurate prediction. This study recommends risk mitigation strategies, such as managing work schedules, increasing rest time, and optimizing workload distribution. In addition, the development of more accurate prediction models with machine learning or multilevel logistic regression is recommended to improve the effectiveness of the analysis.