28 November 2025 in Computer Applications in Engineering Education
This study reports on the design, implementation, and evaluation of an AI‐Supported Pedagogical Architecture (PA) aimed at fostering Self‐Regulated Learning (SRL) within a Virtual Learning Environment (VLE). The architecture was integrated into the Moodle platform and applied in an introductory Python course. It incorporates technological tools such as Completion Progress, Analytics Graphs, Athena, and the Time Tracker SRL plugin to deliver real‐time feedback, facilitate progress monitoring, and support learners in goal setting, planning, and reflection–key dimensions of SRL. Through Educational Data Mining techniques, the study identifies distinct learner profiles based on interaction logs, revealing a significant correlation between SRL strategies and academic achievement. Additionally, qualitative data from focus groups underscore the architecture's effectiveness in enhancing metacognitive awareness, student engagement, and overall satisfaction, demonstrating its potential to support the development of learner autonomy in online education.
28 May 2026 in Revista Brasileira de Informática na Educação
Computers have become an integral part of everyday life. In education, these technologies enable the creation of Ubiquitous Learning Environments, which enrich the learning experience by providing more dynamic and engaging contexts, both in-person and online. Educators and institutions increasingly adopt technological tools to strengthen teaching and learning processes. As a result, the extensive use of these technologies has produced valuable data repositories that can be explored through data mining methods. In this context, this study presents an exploratory analysis of interaction data collected from a Ubiquitous Learning Environment, using Educational Data Mining techniques. Clustering methods were applied to explore students’ behavior in learning sessions and to identify patterns associated with their quiz performance in the environment. Three data mining algorithms were applied to five distinct datasets, each prepared with different preprocessing strategies. The results showed that clustering performance is highly sensitive to data preprocessing, with the best outcomes achieved using feature selection and aggregation techniques. The findings showed statistically significant distinctions among the clusters and uncovered evidence of self-regulated learning within one of the groups. However, the effect size analysis indicated that only a subset of attributes, particularly those related to study time and interaction, presented substantial practical differences between clusters. These results suggest that students exhibiting higher levels of interaction and engagement tend to achieve better performance, indicating behavioral patterns potentially associated with self-regulated learning.