Aristotle University Of Thessaloniki (AUTH)
Τεχνολογίες μάθησης για την υποστήριξη εκπαιδευομένων σε Μαζικά Ανοικτά Διαδικτυακά Μαθήματα (MOOCs)
Abstract
dc:descriptionMassive Open Online Courses (MOOCs) emerged with the promise of making knowledge globally accessible, transcending geographical, economic, and so-cial barriers. However, their educational potential is undermined by the persis-tent problem of extremely high dropout rates. The majority of participants en-roll but do not complete the courses, thereby limiting the social and pedagogi-cal impact of this technology. The dropout problem cannot be reduced merely to a technical issue; it is deeply linked to challenges regarding how learners motivate themselves, organize their study, and manage their time and cognitive resources in highly autonomous environments. At the core of these challenges lies Self-Regulated Learning (SRL). SRL is a critical determinant of success, as it defines the extent to which learners are able to set goals, monitor their progress, adapt strategies, and persevere in the face of difficulties. In MOOCs, where the lack of direct instructional support and the sense of isolation are common, self-regulation skills become even more deci-sive. Learners with low SRL are more likely to drop out quickly, while those with high SRL demonstrate greater resilience, consistency, and likelihood of course completion. Strengthening SRL is therefore not merely desirable but essential for the sustainability of MOOCs. This thesis approaches the dropout problem through three interconnected re-search dimensions. First, it examines the performance of classical machine learning algorithms in predicting dropout, employing different evaluation met-rics and data balancing techniques in order to achieve more reliable estimations of dropout propensity. Second, it highlights the significance of learners’ self-reported data, identifying crucial predictive indicators with particular emphasis on SRL, which proves to be a key factor in both predicting and preventing dropout. Third, it explores the development and implementation of artificial intelligence–based interventions, designed to actively foster learners’ self-regulation and provide personalized guidance in real time. Drawing on empirical studies conducted in real MOOCs, the thesis demonstrates that effectively addressing dropout requires the integration of three strategies: algorithmic prediction, the cultivation of SRL skills, and dynamic support through artificial intelligence tools. Dropout prediction is not an end in itself; it serves as an early diagnostic mechanism. SRL interventions are not merely the-oretical approaches; they function as practical instruments for empowering learners. And artificial intelligence is not a neutral technological addition; it becomes a catalyst enabling personalization, scalability, and continuous adapta-tion of support. Overall, the thesis argues that the future of MOOCs depends on their capacity to evolve from passive content repositories into dynamic learning ecosystems, where self-regulation is actively cultivated and supported, and where artificial intelligence acts as a mediator between dropout risk diagnosis and the en-hancement of learning skills. The proposition is clear: only through an interdis-ciplinary, multifactorial approach can MOOCs overcome their inherent limita-tions and emerge as truly sustainable and effective tools of education.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Psathas, Georgios
- Ψαθάς, Γεώργιος
Subjects
dc:subject × 14- Μαζικά ανοιχτά διαδικτυακά μαθήματα
- Αυτορρύθμιση της μάθησης
- Παρεμβάσεις μάθησης
- Αλγόριθμοι πρόβλεψης εγκατάλειψης
- Massive open online courses (MOOCs)
- Self-Regulated Learning (SRL)
- Learning interventions
- Dropout predictive algorithms
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Φυσικές Επιστήμες
- Επιστήμη ηλεκτρονικών υπολογιστών και Πληροφορική, άλλοι τομείς
- Computer and Information Sciences
- Natural Sciences
- Computer and Information sciences, miscellaneous
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/60363