Predictive Modeling for Detection of Depression Using Machine Learning

Predictive modeling techniques using artificial intelligence have shown promising potential in detecting and predicting depression in recent times, adding newer perspectives to mental health assessment and treatment. This paper presents a predictive modeling approach to detect the presence of depression using machine learning techniques. It presents predictive models to detect depression based on depression-related data of a student cohort containing demographic and academic data along with depression information collected through the Beck Depression Inventory questionnaire, in addition to scores such as the PHQ (Patient Health Questionnaire) score, GAD (Generalized Anxiety Disorder) score, and Epworth score, which provide insights into the severity and impact of depressive symptoms, anxiety symptoms, and daytime sleepiness, respectively. The methodology involves data collection and preparation, feature selection, model selection, and model training using machine learning techniques. The results show the performance metrics of different predictive models on various dataset versions generated through preprocessing steps such as normalization, feature encoding, and selection. The best metrics are compared and evaluated, where the Linear Discriminant Analysis model performed best in terms of AUC, F1 score, and other metrics in this specific cohort. Considering the recent advancements of machine learning, incorporating predictive modeling would be important to designing clinical decision support systems, for a comprehensive prediction and analysis of depression in different cohorts, to act as an assistive tool for mental health professionals.