GCN-BERT Models for Real-Time Occupational Well-Being Assessment of College Teachers in Online Learning Environments

Abstract

In the context of online education, this study introduces a GCN-BERT model to assess college teachers' occupational well-being in web-based learning environments. The model analyzes text data from online teaching platforms, constructing a “teacher-text-semantic” graph to capture dynamic well-being trends. Positive online discussions correlate with 18.7% higher student course completion and 12.3% better assignment quality, while negative feedback increases stress-related semantic weights by 15.2%. Applied in virtual teaching communities and AI tutor systems, the model reduces management workload, improves teacher well-being scores by 11 points, and enhances professional competence by 22.4%. This framework demonstrates the potential of graph neural networks and pre-trained models to transform online faculty management and enhance teaching effectiveness.