University of the Aegean
Υπολογιστικά υποστηριζόμενη εκπαιδευτική αξιολόγηση σε περιβάλλοντα ηλεκτρονικής μάθησης
Abstract
dc:descriptionDigital technologies are transforming education. They are supporting the development and operation of new learning environments, where the processes of educational assessment are provided through computer systems. Computer Assisted Assessment is associated with a wide range of systems and processes. It includes systems and applications that automate traditional assessment methods, but it also includes innovative methods, where assessment acquires new forms and is integrated into the new digital learning environments. The thesis presents in detail the basic forms that Computer Assisted Assessment can take, focusing on its most innovative forms. The aim of the thesis is to contribute to the study of predictive Machine Learning models in digital learning environments and the abilities of their utilization in educational assessment. In particular, the thesis explores the potential of the models to utilize data from Learning Management Systems (LMS) and to predict student performance, not only with accuracy, but also in time, so that they can be used as formative assessment tools, which can identify factors affecting academic failure and guide targeted educational interventions. Data from the LMS Moodle was collected for 4 undergraduate and postgraduate courses. The predictive models were evaluated for accuracy and for timely prediction. Dataset features were identified, which affect the performance of the models to a different degree and which are related to student engagement. The main results of the research showed that, Machine Learning models, predict student’s performance with quite good accuracy and reliability. In addition, early prediction was possible, well before the completion of the courses. It emerged that some of the dataset features affect the prediction performance more than others. These features are common among relevant courses and their high degree of influence can be attributed to their relationship with indicators of student engagement. It also emerged that differences in the structure and design of courses, affect the importance of dataset features and ultimately affect the performance of the models.
Degree
thesis:*- Grantor dc:publisher
- University of the Aegean
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Tsigaros, Theologos
- Τσιγάρος, Θεολόγος
Subjects
dc:subject × 32- Υπολογιστικά Υποστηριζόμενη Αξιολόγηση
- Ψηφιακή Αξιολόγηση
- Ψηφιακή μάθηση
- Μαθησιακή Aναλυτική
- Εξόρυξη Εκπαιδευτικών Δεδομένων
- Προβλεπτική Αναλυτική
- Επιδραστικά Χαρακτηριστικά
- Ενεργή Συμμετοχή
- Computer Assisted Assessment
- e-assessment
- E-learning
- Learning analytics
- Educational data mining
- Prediction Analytics
- Infuential features
- Engagement
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Φυσικές Επιστήμες
- Διεπιστημονικές εφαρμογές επιστήμης ηλεκτρονικών υπολογιστών
- Τεχνητή νοημοσύνη
- Εκπαίδευση
- Κοινωνικές Επιστήμες
- Εκπαίδευση, άλλοι τομείς
- Διδασκαλία και κατάρτιση
- Computer and Information Sciences
- Natural Sciences
- Computer Science Interdisciplinary Applications
- Artificial Intelligence
- Education
- Social Sciences
- Education, miscellaneous
- Teaching and Training
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/53890