Comparison of machine learning models for the identification of disc herniation

The accurate identification of disc herniations is crucial for the proper diagnosis and treatment of low back pain, a health issue affecting millions of people worldwide. Therefore, this study proposes to develop an innovative approach by comparing machine learning models applied to data obtained from a smart belt equipped with advanced sensors. The adopted methodology encompasses data collection, preprocessing, and analysis, evaluating several models, including Logistic Regression, Decision Trees, Naive Bayes, Random Forest, and k-NN. The results reveal that Logistic Regression is the most effective model, demonstrating the best performance metrics in accuracy, AUC, recall, and MCC, with an AUC of 0.763. This model surpasses others in identifying disc herniations and offers balanced and consistent performance. Thus, the utility of machine learning models in clinical practice is validated, proposing a promising tool to improve the diagnosis and treatment of disc herniations. The findings establish a solid foundation for future research, highlighting the importance of exploring this technological approach to contribute to advancements in health and the management of low back pain