What drives the shift? A data mining approach to understanding student preferences for digital learning in post-pandemic Malaysia

Abstract

The rapid evolution of post-pandemic education has highlighted a significant shift toward online platforms. However, the specific emotional factors and core competencies driving university students to choose digital over traditional education remain largely unexplored. Therefore, this study aims to identify the underlying criteria that differentiate traditional from digital learning and determine the key drivers of students' shifting preferences. Methodologically, a quantitative data mining approach was employed, using secondary data from the Department of Statistics Malaysia, comprising data on 432 university students. The data were analyzed in Orange software using linear regression and the RReliefF feature selection method to extract predictive patterns. The results indicate that students distinguish between the two learning modes primarily based on their study skills, feelings of academic overload, and a lack of interaction. Furthermore, a preference for digital education is significantly influenced by personal interest, problem-solving abilities, and managerial/entrepreneurial skills. In conclusion, these data-driven insights emphasize that successful digital learning environments must actively support student well-being and interactive engagement. This study equips educational policymakers and instructional designers with the targeted data needed to optimize future educational environments.