Role of Individual and Occupational Factors on Injuries among Contractual Workers in Surface Mines-Machine Learning Approach

This study integrates data analysis and machine learning techniques to investigate workplace injury prediction and mitigation in surface mining operations. Injury data (from March 2023 to Aug 2024) was collected from mechanised limestone mines in Raipur, Chhattisgarh, encompassing 446 workers' demographic details, health data and workplace conditions. The methodology involved exploratory data analysis, linear and non-linear regression models and advanced machine learning algorithms such as XG Boost, random forest, decision tree, and Bayesian ridge regression. risk factors namely skill level (semiskilled), age (mean 43 years), experiences (mean 14 years), designation, qualifications, health status, work environment, and safety culture were analysed to predict injury occurrences. The results reveal that near misses and minor incidents are significant early indicators (T-value 1.35) of severe injuries, emphasising proactive safety measures. Among the models evaluated, XG Boost well performed with a training R-squared of 0.95 and test R-squared of 0.43, demonstrating its superior ability to generalise and capture complex relationships in the data. The analysis found that injuries are moderately common for contractual workers (mean: 0.90 per instance), with significant variability in near misses (mean: 8.05). Standardised working conditions, such as uniform duty hours (8 hrs.) and leave entitlements (30 days/year), contribute to the occurrence of injuries (t-value 1.35). Proactive measures to be taken in addressing factors like worker skill levels, availability of safety features and positive attitudes toward safety, are crucial for mitigating injuries at workplaces. Major Findings: From the study it can be concluded that the factors namely fire (T-value=5.64), experience (T-value=4.69), skill level (T-value=5.08), positive attitude towards safety (T-value=3.68), medical facilities (T-value=4.35), working conditions of the site (T-value=5.23) and available equipment with safety features (T-value=4.55) are the significant factors that contribute to the occurrence of the injuries among contractual workers in surface mines. The machine learning algorithm models namely extreme gradient boost (XG Boost) performed best for the prediction of the injuries data consider for the study with training data set, (R-squared 0.95) and testing data set (R-squared 0.43).