Learning analytics in online education: data-driven insights into student success

Abstract

In order to gain insights into the key factors affecting students' success in online education, this paper extracts students' online learning behavioral feature indicators through the behavioral record data in the online learning platform, applies the attribute approximation algorithm based on the Bayesian Fuzzy Rough Set (IDB-BRS) model to attribute approximation of the behavioral indicators, and utilizes the improved Apriori algorithm to mine the association rules between online learning behaviors and learning effects. The improved Apriori algorithm is used to establish association rules between online learning behaviors and learning effects. In comparison to the VPFRS model attribute approximation algorithm, the IDB-BRS model attribute approximation algorithm does not necessitate pre-given parameters and achieves superior classification accuracy and approximation time in the Soybean, Credit, and Balance datasets, thereby offering greater practical value. The association rules reveal that students who carefully study course resources, actively submit assignments, and study online frequently contribute positively to their success in online learning. This paper holds significant implications for enhancing the effectiveness of learning in online education.