Classification of Depression and Anxiety with Machine Learning Applying Random Forest Models

Anxiety and depression have a strong correlation; although they often differ in their symptoms, both disorders frequently coexist in individuals, leading to a significant impairment of their daily functioning. Anxiety is among the most common conditions that, over time, can develop into a depressive disorder, or vice versa. The simultaneous presence of these disorders contributes to a range of symptoms, including mood disturbances, anhedonia, sleep problems, persistent fear, suicidal thoughts or ideation, distress, irritability, and constant worry, which result in substantial suffering for affected individuals. Recognizing this complex interplay, the present research applied Machine Learning models, particularly Random Forest, which is renowned for its robustness in classification tasks and its ability to identify both linear and non-linear relationships. The study involved the use of 17 predictor variables to train two models, each focused on predicting different outcomes: “Depression Severity” and “Anxiety Severity.” By leveraging the strengths of Random Forest, the research achieved highly accurate classifications. Specifically, the model for “Depression Severity” attained an accuracy of 99.35% and an F1 Score of 99.34%, while the model for “Anxiety Severity” reached an accuracy of 98.04% and an F1 Score of 98.05%. These results underscore the effectiveness and precision of the Random Forest algorithm in accurately identifying the severity of these mental health conditions.