DEPRESSION DETECTION USING MACHINE LEARNING ALGORITHMS: A SYSTEMATIC LITERATURE REVIEW

Abstract

Depression is a common mental health challenge significantly affecting the global health community. Recently, researchers have conducted studies on applying machine learning methods for automated identification of depression symptoms among victims. Depression diagnosis now uses machine learning approaches to identify depression in social media posts. With these machine-learning techniques, researchers improve the detection and prediction of health-related diseases to support government and medical health professionals in their decision-making process. This paper aims to review and assess machine learning methods for the detection of depression. The present prediction systems used by the researchers meticulously found victims of depression using identification, detection, analysis, and prediction. Boolean keywords were used to search several journal databases and filters; the researchers chose and assessed thirty (30) papers, emphasising the application of machine learning methods. The results show that many studies used different models, such as SVM, RF, DT, K-NN, and NB, with various datasets to identify depression in people using clinical or social media information. Datasets utilised in the papers reviewed include those from Twitter, Facebook, Reddit, Survey, and DASS 21. There were few studies on prediction 17%, but most focused on early depression detection. The analysis showed a decline from 2022 to 2024, but the highest rise in machine learning research was in 2019.