Un recomendador para ayudar en la evaluación de la participación en foros de aulas virtuales
- Javier Luis Cánovas Izquierdo 1
- José Antonio Hernández López 2
- Jesús Sánchez Cuadrado 2
- David Bañeres Besora 1
- 1 Universitat Oberta de Catalunya (UOC)
-
2
Universidad de Murcia
info
- Cruz Lemus, José Antonio (coord.)
- Medina Medina, Nuria (coord.)
- Rodríguez Fortiz, María José (coord.)
ISSN: 2531-0607
Year of publication: 2023
Issue Title: Actas de las XXIX Jornadas sobre la Enseñanza Universitaria de la Informática. Granada, del 5 al 7 de julio de 2023
Issue: 8
Pages: 351-358
Type: Article
More publications in: Actas de las Jornadas sobre la Enseñanza Universitaria de la Informática (JENUI)
Abstract
Virtual communication channels such as discusión forums are offered in many subjects of bachelor’s degrees. The students’ participation in these spaces allows the assessment of skills such as written communication, argumentation ability or collaborative work, among others. Thus, many subjects propose activities where students have to work collaboratively, discuss and reach agreements to develop a solution. This discussion can be represented as collaboration graphs that allow visualising the discursive environment of the forum. In this work, we present a solution based on Machine Learning techniques that recommends an assessment mark of the students’ participation in virtual discussion forums based on collaboration graphs, thus helping the teaching staff in the evaluation process. We analyze two machine learning architectures: models using graph metrics and graph neural networks, being the latter the one that offers the lowest error rate.