Educational Data Mining (EDM) plays a pivotal role in supporting evidence-based decision-making in online and distance education. Yet, many EDM models lack semantic transparency, which limits their interpretability and adaptability in dynamic learning environments. In this session, I will introduce a semantic approach that integrates ontologies and Semantic Web Rule Language (SWRL) with machine learning models - particularly decision trees - to enhance the prediction of student performance. Drawing from my recent research published in The International Review of Research in Open and Distributed Learning (IRRODL), I will demonstrate how ontology-based reasoning can complement EDM to create intelligent, adaptive systems that support both learners and institutional interventions. The session will also offer a broader introduction to EDM as a research methodology, including its potential applications, key challenges, and future directions.