Optimizing Naïve Bayes Method for Felder-Silverman Learning Style Model Identification

Abstract

One important issue in education institusion is the differences in students learning styles, which requires educators to pay attention to individual learning preferences. The manual learning style identification method is considered less effective in terms of time and data accuracy. This study aims to develop a student learning style identification system using the Felder-Silverman model and the Naïve Bayes method, This system is designed to assist lecturers in adjusting learning strategies according to student learning preferences, thus increasing the effectiveness of the learning process. The Naïve Bayes method was applied by analyzing student datasets and determining the accuracy of learning style identification. The validation results showed significant identification accuracy: 85% for the active-reflective dimension, 96% for the sensitive-intuitive dimension, 98% for the verbal-visual dimension, and 91% for the sequential-global dimension. The results of user validation show the effectiveness of the learning style identification application that has been tested based on the percentage value of each statement, and an average percentage value of 85.6% was obtained for all statements, indicating that the system functions well in identifying students' learning styles, while the results of expert validation state that the statements are in accordance with the indicators, the statements use simple and easy-to-understand language, and the identification results are appropriate. This study is expected to contribute to helping universities identify student learning styles efficiently, improve the quality of learning in higher education, and contribute to supporting an inclusive learning approach in higher education environments.