ENSEMBLE LEARNING AND ITS RELATIONSHIP WITH INTERPRETABILITY FOR THE DETECTION OF DEPRESSION IN CHILDREN AND ADOLESCENTS

This study addresses the early diagnosis of depression in young people aged 10 to 16 years, using machine learning techniques. Five algorithms were applied, including individual and ensemble methods. The objective of this study was to evaluate the performance of these models in relation to the metrics traditionally known as precision sensitivity, in addition to the discussion of these methods in terms of interpretability. The results revealed that the ensemble methods, especially Random Forest and XGBoost, showed better performance. However, for these models there is a greater need for an investigation if the objective is to identify the important variables in the context of depression. This discussion highlights the potential of advanced computational approaches to complement traditional clinical practices in healthcare.