Machine Learning Approaches for Social Media Based Depression Detection: A Review

Abstract

Depression is one of the most prevalent mental health disorders that affect millions of individuals and leading to severe psychological, social and economic consequences. Traditional diagnosis of depression relies mainly on clinical interviews and psychological assessments. However, recent advances in machine learning, natural language processing (NLP) and physiological signal analysis have enabled automated detection of depression using social media data and bio-signals. This review summarizes current computational approaches for depression detection, focusing on machine learning models with textual and physiological data sources. It also discuss data sources, feature extraction techniques and algorithms used in depression detection.