EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
3491. A network-based recommender system to enhance learning in educational games
Invited abstract in session WA-27: Learning Analytics using Mathematical Optimization and XAI, stream Mathematical Optimization for XAI.
Wednesday, 8:30-10:00Room: 047 (building: 208)
Authors (first author is the speaker)
1. | Eva Ósk Gunnarsdóttir
|
Department of Computer Science, Reykjavík University | |
2. | Anna Ingólfsdóttir
|
Department of Computer Science, Reykjavík University | |
3. | Antonio Puron
|
Inoma / Tak-Tak-Tak | |
4. | María Óskarsdóttir
|
Department of Computer Science, Reykjavik University | |
5. | Sebastian Maldonado
|
Department of Management Control and Information Systems, University of Chile |
Abstract
Educational games hold particular significance in elementary school education, as they serve as a popular tool for engaging children in learning. However, a common challenge faced by these platforms is the risk of children becoming disinterested. To mitigate this, it becomes desirable to sustain their engagement with the platform to foster continuous learning and retention. This raises the crucial question: Which games should children play? While the approach may be effective in the classroom, children who engage with these games outside of school settings have unrestricted access to all available games, often without knowing which ones are best suited to their needs. Consequently, if they struggle, they are prone to becoming bored or frustrated, ultimately leading them to disengage. To enhance engagement and optimize learning outcomes, it is essential for children to play games where they perform adequately yet are still challenged enough to promote learning. In this work, we develop state-of-the-art recommender systems based on graph-based neural networks for an educational game platform for children called Tak-Tak-Tak. The machine learning process is guided by the choices made by the users in terms of the next game to play, as well as the efficacy of a user's learning experience within a game and the duration of the play sessions. We conduct a comparative analysis between a wide variety of recommender system approaches to evaluate their effectiveness in this context.
Keywords
- Analytics and Data Science
- Machine Learning
- OR in Education
Status: accepted
Back to the list of papers