<![CDATA[This study aims to analyze and classify user satisfaction review texts of the Faculty of Science and Technology (FST) Library into three categories: Dissatisfied, Satisfied, and Very Satisfied. The study used 1,096 library service survey reviews collected between 2023 and 2025 as the dataset for analysis. The research process comprises text preprocessing, such as stemming, stopword removal, tokenizing, case folding, and cleaning, and filtering), The study applied TF-IDF weighting, divided the dataset into 70:30 training and testing sets, and performed clustering with the K-Means algorithm into three clusters. The resulting clusters were assigned to satisfaction categories and evaluated using an accuracy, precision, recall, and F1-score confusion matrix. The model achieved 87.23% accuracy, 86.67% weighted precision, 87.23% weighted recall, and 86.85% weighted F1-score. These results showed that the K-Means algorithm can be used as a foundation for assessing and enhancing the quality of library services and is capable of efficiently classifying user satisfaction reviews. ]]>
Copyrights © 2026
