<![CDATA[A Detection or identification of student learning styles has a significant role in improving the teaching and learning process. Teachers who are aware of the diverse learning styles of their students can make learning adaptations or prepare learning instructions appropriately, while students who are aware of their learning styles can adapt their learning methods so as to produce better learning process outcomes. However, detecting student learning styles is complex and challenging and there are various factors that can affect the accuracy of detection. Various literature studies on detecting student learning styles note that there are various problems that occur in Virtual Based Learning Education (VBLE), namely the lack of investigations regarding the suitability of teaching and learning styles making it difficult to increase student learning motivation, the traditional instructions of some teachers which are still in conflict with the student learning style system so that it is difficult to adapt to the learning environment and most students do not know effective learning and do not pay the necessary attention to their learning style. This study aims at two (2) things, the first is to detect the learning styles of private high school students (SMAS) in VBLE with the Decision Tree (DT) C4.5 and Naive Bayes (NB) algorithms and the second is to compare the performance of DT and NB algorithms for the context of style detection. private high school students study on VBLE. Data collection was carried out using a google form questionnaire on 252 respondonce consist of alumni and students through 30 questions about learning styles. Student learning styles are identified as visual, auditory and kinesthetic learning styles. For data analysis, the Google Colab tool is used with Python programming. While the process of evaluating the performance of the algorithm is measured using the parameters accuracy, precision, recall/sensitivity, f-measure and time consumption. The results showed that the DT and NB algorithms were good at predicting SMAS learning styles, but the NB algorithm (98%) was slightly better than the DT algorithm (96%) in terms of accuracy, precision, memory/sensitivity, f-measure, while for time consumption DT (0,00505 seconds) much faster than NB(3,2895 seconds). This research contributes in two (2) ways, first scientifically by testing the DT and NB algorithms for detecting learning styles of SMAS students in the VBLE context and practically providing recommendations to education stakeholders namely teachers, students, student guardians and education policymakers to further pay attention to the detection of student learning styles]]>
