Depression is one of the most common mental health disorders worldwide and has become a growing concern in modern healthcare systems. Early identification of depression risk factors can significantly improve prevention strategies and clinical intervention planning. This study proposes a machine learning–based framework for predicting depression risk using behavioral and demographic indicators. Multiple machine learning models, including Logistic Regression, Random Forest, and XGBoost, were developed and evaluated to analyze their predictive performance. Experimental results demonstrate that ensemble learning models outperform traditional classification methods in depression prediction tasks. Model evaluation using confusion matrix, ROC analysis, and precision–recall metrics indicates strong predictive performance of the proposed models. Feature importance and SHAP-based explainable AI analysis further reveal that sleep patterns and stress levels are the most influential predictors of depression risk. The findings highlight the potential of machine learning–based predictive analytics for early depression detection and mental health monitoring.
Dipta Paul (Tue,) studied this question.