A psychological education model integrating artificial intelligence-based BERT in physical education

Area of Science:

• Sports Psychology
• Natural Language Processing
• Computational Linguistics

Background:

• Traditional physical education (PE) relies on manual observation for psychological assessment, which has limitations.
• Integrating Natural Language Processing (NLP) offers a more objective and scalable approach to psychological evaluation in educational settings.

Purpose of the Study:

• To develop an intelligent psychological education model using NLP and the Bidirectional Encoder Representations from Transformers (BERT) model for precise student psychological state evaluation in PE.
• To overcome the limitations of manual observation in traditional PE psychological assessments.

Main Methods:

• A text sentiment analysis model was built to extract emotional features.
• The BERT model was fine-tuned for sports-specific contexts to capture deep semantic features.
• A four-dimensional feature extraction framework (semantic, emotional, thematic, psychological) was established.

Main Results:

• The BERT-LDA-Psych model achieved high accuracy (91.5%), precision (91.0%), recall (90.3%), and F1 scores (90.6%) on sentiment analysis tasks.
• The model outperformed mainstream ensemble models in psychological state recognition and feature fusion efficiency.
• Stable performance was observed on general psychological counseling text corpora.

Conclusions:

• The proposed BERT-LDA-Psych model provides a feasible computational framework for psychological state recognition in PE contexts.
• While effective in controlled environments, further research is needed to assess transferability to authentic PE classroom discourse.