Scholarlink

Background: Work-related musculoskeletal disorders (WRMSDs) are widespread among mining machine operators due to prolonged whole-body vibration (WBV) exposure and ergonomic risks. A reliable predictive model is essential to accurately estimate WRMSD prevalence and support preventive strategies. Methodology: This study evaluated the predictive performance of four machine learning models—Random Forest, Support Vector Machine (SVM), XGBoost, and Artificial Neural Network (ANN)—for identifying WRMSD risk among shuttle car operators. Data were collected from 54 operators (mean ± SD: age = 40.38 ± 7.19 years; BMI = 24.98 ± 1.99 kg/m2; experience = 12.06 ± 3.14 years). Key input variables included age, BMI, experience, posture (REBA score), frequency-weighted RMS acceleration, and vibration dose value (VDV). WRMSDs prevalence and severity were assessed using the standardized Nordic Musculoskeletal Questionnaire. Results: Significant correlations were observed between WRMSD severity and age, experience, posture, and vibration exposure [A(8), VDV(8)]; BMI was not significantly associated. Among the models, ANN achieved the highest performance (AUC-ROC: 0.9688; F1-score: 0.9143; Recall: 1.0000). Conclusion: The ANN model demonstrated superior accuracy in predicting WRMSDs and holds strong potential for proactive risk assessment in mining environments. Further model refinement and validation across diverse settings are recommended for broader occupational health applications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicting Musculoskeletal Disorders Among Shuttle Car Operators Using Machine Learning Techniques

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ali Murtoja Shaikh,
  • Bibhuti Bhusan Mandal

摘要

Background: Work-related musculoskeletal disorders (WRMSDs) are widespread among mining machine operators due to prolonged whole-body vibration (WBV) exposure and ergonomic risks. A reliable predictive model is essential to accurately estimate WRMSD prevalence and support preventive strategies. Methodology: This study evaluated the predictive performance of four machine learning models—Random Forest, Support Vector Machine (SVM), XGBoost, and Artificial Neural Network (ANN)—for identifying WRMSD risk among shuttle car operators. Data were collected from 54 operators (mean ± SD: age = 40.38 ± 7.19 years; BMI = 24.98 ± 1.99 kg/m2; experience = 12.06 ± 3.14 years). Key input variables included age, BMI, experience, posture (REBA score), frequency-weighted RMS acceleration, and vibration dose value (VDV). WRMSDs prevalence and severity were assessed using the standardized Nordic Musculoskeletal Questionnaire. Results: Significant correlations were observed between WRMSD severity and age, experience, posture, and vibration exposure [A(8), VDV(8)]; BMI was not significantly associated. Among the models, ANN achieved the highest performance (AUC-ROC: 0.9688; F1-score: 0.9143; Recall: 1.0000). Conclusion: The ANN model demonstrated superior accuracy in predicting WRMSDs and holds strong potential for proactive risk assessment in mining environments. Further model refinement and validation across diverse settings are recommended for broader occupational health applications.