This research introduces an LLM-based framework that better identifies mental health crises on social media by integrating psychological health expertise into language models.

Key Innovations:

• Multi-level framework combining BERT transfer learning with mental health domain knowledge
• Integration of sentiment analysis and behavioral prediction capabilities
• Specialized approach for detecting crisis signals in social media conversations

Business Impact: This methodology enables more accurate and timely identification of mental health crises on social platforms, potentially allowing healthcare providers and platform moderators to intervene earlier and more effectively with at-risk individuals.

Psychological Health Knowledge-Enhanced LLM-based Social Network Crisis Intervention Text Transfer Recognition Method