Predicting Student Academic Performance using Data Generated in Higher Educational Institutes

  • Areej Fatemah Meghji
  • Naeem Ahmed Mahoto
  • Mukhtiar Ali Unar
  • Muhammad Akram Shaikh


The analysis of data generated by higher educational institutes has the potential of revealing interesting facets of student learning behavior. Classification is a popularly explored area in Educational Data Mining for predicting student performance. Using student behavioral data, this study compares the performance of a broad range of classification techniques in an attempt to find a qualitative model for the prediction of student performance. Rebalancing of data has also been explored to verify if it leads to the creation of better classification models. The experimental results, validated using well-established evaluation matrices, presented potentially significant outcomes, which may be used for reshaping the learning paradigm.