An Analysis of Self-Regulated Learning Behavioral Diversity in Different Scenarios in Distance Learning Courses

Abstract

The increasing volume of student behavioral data within virtual learning environments (VLE) provides many opportunities for knowledge discovery. Thus, techniques that make it possible to predict the academic performance of students become essential tools to assist distance learning instructors. This article shows the results of the development of a student performance predictive model, based on behavioral indicators of self-regulated learning in a database extracted from the Moodle VLE. In addition, we attempted to develop specialized predictive models for three distinct scenarios (general, divided by course and divided by semester). The results showed that the variation in the student behavior through the different semesters has a strong influence on the model's predictive power.