Psychological Disorder Identifying Method Based on Emotion Perception over Social Networks

Area of Science:

• Computational psychiatry
• Natural Language Processing (NLP)
• Social media analytics

Background:

• Global mental disorders affect over 1.1 billion people, impacting social development.
• Challenges in timely diagnosis hinder effective mental health care delivery.
• Existing mental health action plans require enhanced diagnostic tools.

Purpose of the Study:

• To develop a mental disorder-aided diagnosis model (MDAD) for analyzing user-generated text on social networks.
• To quantify multipolarity sentiment affect intensity and analyze 11-dimensional sentiment distribution.
• To enable real-time detection of individuals at high risk for mental disorders.

Main Methods:

• Data collection from Twitter hashtags related to five mental disorders.
• Quantification of multipolarity sentiment affect intensity in short texts.
• Sentiment distribution similarity calculations and Stochastic Gradient Descent (SGD) for detection.

Main Results:

• The MDAD model effectively diagnoses common mental disorders including depression, anxiety, OCD, bipolar, and panic disorder.
• Real-time detection of individuals with high probability of mental disorders is achieved.
• Timely mental health warnings are issued for users exhibiting distinct emotional tendencies.

Conclusions:

• The proposed MDAD model offers effective mental disorder diagnosis through sentiment multipolarity analysis.
• This approach provides strong support for the prevention and early diagnosis of mental disorders.
• Sentiment analysis of social media data can be a valuable tool in public mental health surveillance.