<![CDATA[This study aims to evaluate students’ satisfaction with academic services and analyze open-ended opinions using the Artificial Intelligence Bidirectional Encoder Representations from Transformers (AI-BERT) model. The research employed a quantitative experimental method, combining a five-point Likert scale survey across seven academic service indicators and AI-BERT sentiment analysis of 150 student comments. The results indicate that the overall student satisfaction level falls into the “good” category, with a Student Satisfaction Index (SSI) of 78.53%. The highest-rated indicator was access to academic information (mean = 4.21), while the lowest was administrative service speed (mean = 3.67). Sentiment analysis revealed 60.67% positive, 36.00% negative, and 3.33% neutral opinions, highlighting the need for improvement in service speed and staff responsiveness. Evaluation of the AI-BERT model demonstrated superior performance with an accuracy of 91.3% and an F1-score of 0.913, outperforming conventional methods such as SVM and Naïve Bayes. These findings provide a basis for recommendations on developing digital-based academic service strategies and leveraging AI technology to enhance service efficiency and quality.]]>
Copyrights © 2025
