Detecting Mental Health Disorders Through Emotional Patterns on Social Media: The Case of Anorexia

The objective of this study is to identify anorexia symptoms on social media by examining user postings' emotional patterns. The model utilizes sentiment analysis and Natural Language Processing (NLP) with algorithms like Bidirectional Encoder Representations from Transformers (BERT) to identify emotional signs of anorexia, including depression, anxiety, and problems with self-esteem. A Long Short-Term Memory (LSTM) neural network detects persistently unfavorable emotional states by further analyzing temporal patterns. To provide an approach for early mental health issue identification and prompt intervention, the system is intended to monitor and flag posts that could possibly indicate anorexia risk.