Predicting Students Academic Performance Based on Learning Analytics

Abstract

In the context of the digital transformation of higher education, Learning Analytics has emerged as an important approach for predicting students’ academic performance. Traditional assessment methods often do not fully capture the complex relationships between learners’ behavioral, cognitive, and motivational factors. This study develops a structural model for predicting academic achievement using Learning Management System (LMS) data and Structural Equation Modeling (SEM). The research involved 850 university students studying in a blended learning environment. The analysis included digital engagement indicators such as LMS login frequency, forum participation, time spent on learning materials, assignment completion rates, and formative assessment results. Confirmatory Factor Analysis (CFA) was conducted, and reliability and validity were assessed using Cronbach’s alpha, Composite Reliability (CR), and Average Variance Extracted (AVE). The structural model demonstrated satisfactory fit indices (CFI = 0.94, TLI = 0.92, RMSEA = 0.048, SRMR = 0.041). The results showed that digital engagement has an indirect effect on academic performance through the mediating role of academic self-regulation. The proposed SEM model confirms the predictive value of Learning Analytics and reveals the mechanisms linking students’ digital behavior with educational outcomes. The findings support the integration of Learning Analytics into early warning systems and intelligent decision- support tools for identifying academic risks and improving learning processes in higher education.