A Systematic Review of Machine Learning Approaches for Mental Disorder Prediction on Social Media

Mental disorder is a prevalent yet overlooked illness. Social media offers a way to detect mental patients in a broad population. To detect mental patients on social media, many efforts have been taken to analyze individual posts using machine learning techniques. The purpose of this paper is to provide a systematic overview of studies in the mental disorder detection field. A comprehensive review on a list of datasets extracted from various social media platforms and a summary of the types of features explored in mental disorder detection are given in this paper. Also, different regular machine learning techniques and deep learning approaches used in previous studies are discussed. Several directions for future research are also presented in this paper.