The data visualization and intelligent text analysis for effective evaluation of English language teaching

Area of Science:

• Educational Technology
• Natural Language Processing
• Data Science

Background:

• Traditional English language teaching evaluations often rely on subjective teacher judgment and lack real-time monitoring.
• These limitations hinder personalized learning and accurate assessment of student progress.

Purpose of the Study:

• To develop an innovative evaluation system for English language teaching effectiveness.
• To integrate data visualization and intelligent text analysis for objective and personalized educational assessment.

Main Methods:

• Utilized data visualization to convert student learning data into intuitive charts.
• Employed intelligent text analysis, including Bidirectional Encoder Representations from Transformers (BERT) for emotion analysis, Latent Dirichlet Allocation (LDA) for theme modeling, and Word2Vec for semantic analysis.
• Applied ensemble learning methods like Bagging, XGBoost, and Stacking for model optimization.

Main Results:

• The Stacking ensemble model achieved high accuracy rates of 95.0% on the training set and 94.3% on the testing set.
• An optimized model combining BERT, Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN), and self-attention mechanisms demonstrated significant advantages in analyzing student emotions and teaching content.
• The proposed system significantly improves upon single-model performance.

Conclusions:

• The developed system offers a novel approach to educational evaluation, moving beyond subjective traditional methods.
• It provides a powerful tool for achieving personalized and accurate English language teaching and assessment.
• The integration of advanced AI and data visualization enhances the objectivity and effectiveness of educational monitoring.