TRACKING MENTAL HEALTH THROUGH EMOTION ANALYTICS: THE CASE OF ANOREXIA AND DEPRESSION ON SOCIAL MEDIA

Abstract

Social media profiles are becoming more visible indicators of mental health issues including depression and anorexia, which presents new possibilities for the early diagnosis and treatment of these conditions. This research delves at the potential of emotion analytics to spot patterns in language and behaviour linked to these disorders. This study uses natural language processing (NLP), sentiment analysis, and emotion classification methods to sift through massive amounts of usergenerated content on sites like Reddit and Twitter in search of emotional states, sentiment trends, and psychological indicators that point to depression and anorexia. Sadness, fear, rage, and despair are frequent emotions associated with mental health issues, and this technique uses machine learning models trained on annotated datasets to identify them. In order to detect changes in emotional tone over time—a symptom of worsening mental health—temporal and linguistic patterns are also studied. The analytical framework incorporates case-specific elements, such as terminology pertaining to isolation for depression and allusions to body image for anorexia. When paired with contextual behavioural markers, the findings show that emotion-based indications greatly increase the identification of individuals who might be at danger. An effective and scalable technique for monitoring mental health trends using emotion-centric models is proposed in this study, which adds to the fields of computational psychiatry and social media surveillance. Furthermore, it brings attention to the moral questions of privacy, permission, and the appropriate use of AI to delicate fields.