<![CDATA[The Academic Information System supports educational institutions by providing quality management programs to all students and stakeholders, relying on information and communication technologies such as the internet and local networks. Over time, this application has been consistently used by students and lecturers. However, the university has not yet evaluated the feasibility and effectiveness of the system, making it difficult to plan future improvements. Therefore, feedback from both students and lecturers is essential for guiding the system’s development. Given the current state of the application, such evaluation can be carried out through user comments. This study investigates the performance of the Naïve Bayes algorithm, one of the most commonly used algorithms in various research libraries, in analyzing sentiment from these comments. To further improve the accuracy of the Naïve Bayes method, we applied an additional Particle Swarm Optimization (PSO) feature selection process. The results demonstrate that the Naïve Bayes method with PSO achieved an accuracy of 86.27%, precision of 84.78%, and recall of 84.78%, which are higher than the results obtained using the standard Naïve Bayes method alone.]]>
Copyrights © 2025
