An Educational Data Intelligence Model for Early Identification of Slow Learners in BCA Programs Using Multi-Factor Analysis and Explainable Machine Learning

Abstract

 In recent years, higher education institutions have witnessed an unprecedented growth in educational data due to increased student enrollment, digital learning platforms, Learning Management Systems (LMS), and continuous assessment mechanisms. Technology-oriented programs such as the Bachelor of Computer Applications (BCA) generate multidimensional data related to students’ academic, behavioral, psychological, and technical performance. However, traditional evaluation systems rely on delayed assessments, leading to late identification of learning difficulties and ineffective interventions. This paper proposes an Educational Data Intelligence (EDI) model for the early identification of slow learners in BCA programs using multi-factor analysis and explainable machine learning. Primary data collected from 2,028 BCA students through a structured questionnaire were analyzed. After correlation analysis, chi-square validation, and expert review, 20 significant educational intelligence factors were identified. An interpretable decision-tree-based REPTree algorithm was employed to classify students into risk categories. The proposed EDI model provides transparent, actionable insights for early intervention, academic planning, and student retention. The study demonstrates that EDI functions not only as a predictive model but also as an effective academic decision-support system.