Educational Data Mining-based visualization and early prediction of student performance: a synergistic approach

Abstract

Educational Data Mining tools exploit the data automatically logged in e-learning platforms such as moodle to address a variety of education-related questions including the (early) prediction and visualization of student participation and performance data. Efficient tools are available to these ends; they offer a wide spectrum of functionalities at the cost of certain limitations or constraints. A synergistic approach that uses more than one tool is therefore proposed in this paper to answer the double need of the teacher for visualization of current and prediction of future student participation and performance outcomes. Moreover, the feasibility of early and accurate prediction is investigated through a case study on a postgraduate course semester module, aiming at an early warning system for students at risk of failure. Case study results indicate that timely and accurate prediction is possible early enough in the semester to allow enough time for supportive measures within the same term. The combined use of tools can also offer a multi-faceted visualization that dynamically monitors variables of interest at the individual and the class level, including current performance data that feed prediction. More research is certainly necessary, however, to offer to the teachers – end users tools with truly user-friendly interfaces.