Fuzzy Hoeffding Decision Trees for Incremental and Interpretable Predictions of Students' Outcomes

In the last years, also thanks to the spreading of the COVID-19 pandemic, distance learning and the usage of Virtual Learning Environments (VLEs) have experienced a steep increase, becoming powerful tools to support higher education throughout the world. Artificial Intelligence (AI) methods, capable to analyze streams of data (such as logs), can be effectively employed to extract knowledge from them, being useful for all stakeholders involved in the learning process, especially students and teachers. In this abstract, we summarize the results obtained by two stream-based classifiers, namely Hoeffding Decision Tree (HDT) and its fuzzified version FHDT, to predict the students' outcomes in sequential semesters. Moreover, a feature analysis suggesting the most discriminant features for the predictive task has been discussed to explain the reasons behind the success (or failure) of given students in the regarded semesters.