Machine Learning Model to Identify Lumbar Posture in Mine Workers with XGBoost

Joseph Watson Chávez A.; Espinoza B.S.A.; Ovalle C. · Proceedings - 2024 2nd International Conference on Information Network and Computer Communications, INCC 2024 (2024) · DOI: 10.1109/incc64392.2024.00014

Resumen

In this paper, the development of a Machine Learning Predictive Model based on biomechanical data to identify the correct lumbar posture in mining workers is presented. This study is relevant for the scientific and academic community, since it addresses a critical problem in occupational health, where low back disorders are one of the main causes of occupational disability. The approach focuses exclusively on numerical biomechanical values as input for the model, dispensing with the use of image processing or computer vision, which facilitates its practical implementation. The CRISP-DM methodology is adopted, highlighting the phases of preparation, data splitting and fitting, model training and implementation in Gradio. The XGBOOST algorithm is chosen for its efficiency in classifying complex data sets. The training results show a high accuracy of 90%, recall of 89% and F1-score of 89%, indicating that the model can be a valuable tool for early identification of low back problems. This work underscores the importance of addressing occupational health in mining and suggests expansion of the dataset and exploration of new techniques to improve the generalization of the model in future research.

Instituciones: Universidad Privada del Norte