Looking for the best data fusion model in Smart Learning Environments for detecting at risk university students

This paper proposes to discover which data fusion approach and classification algorithm produced the best results from smart classrooms data, and how useful would be the prediction models for detecting University students at risk of failing or dropout. The results showed that the best predictions were produced using ensembles and selecting the best attributes approach with discretized data; the REPTree algorithm demonstrated the highest prediction values. The best predictions also show the teacher what set of attributes and values are the most important for predicting student performance, such as the level of attention in theory classes, scores in Moodle quizzes and the level of activity in Moodle forums.