Optimized machine learning approach may improve understanding of medial tibial stress syndrome in military male personnel

This study investigates the applicability of optimized machine learning approach for the prediction of MTSS using anatomic and anthropometric predictors. To this end, 180 recruits were enrolled in a cross-sectional study (30 MTSS (30.36 ± 4.80 years) and 150 normal (29.70 ± 3.81 years)). Twenty-five predictors/features including demographic, anatomic, and anthropometric variables were selected as risk factors. Bayesian optimization method was used to evaluate the most applicable machine learning algorithm with tuned hyperparameters on the training data. Three experiments were performed to handle the imbalances in the data set. The validation criteria were accuracy, sensitivity, and specificity. The highest performance (even 100%) observed for the Ensemble and SVM classification models while using at least 6 and 10 most important predictors in experiments 1 and 3 respectively. In experiment 2 the best performance (accuracy = 88.89%, sensitivity = 66.67% and specificity = 95.24%) was achieved for the Naive Bayes classifier with 10 most important features. The Naive Bayes, Ensemble, and SVM methods could be the primary choices to apply machine learning approach in MTSS risk prediction. These predictive methods alongside the eight common proposed predictors might help to more accurately calculate the individual risk of developing MTSS at the point of care.