Artificial Intelligence to Counterweight the Effect of COVID-19 on Learning in a Sustainable Environment

Abstract

Distance learning has been adopted as a very extended model during COVID-19-related confinement. It is also a methodology that can be applied in environments where people do not have easy access to schools. In this study, we automatically classify students as a function of their performance and we describe the best self-learning methodologies in distance learning, which will be useful both in confinement or for people with difficult access to schools. Due to the different learning scenarios provided by the different confinement conditions in the COVID-19 pandemic, we have performed the classification considering data before, during, and after COVID-19 confinement. Using a field experiment of 396 students, we have described the temporal evolution of students during all courses from 2016/2017 to 2020/2021. We have found that data obtained in the last month before the final exam of the subject include the most relevant information for a correct detection of students at risk of failure. On the other hand, students who obtain high scores are much easier to identify. Finally, we have concluded that the distance learning applied in COVID-19 confinement changed not only teaching strategies but also students’ strategies when learning autonomously.

Keywords

• supervised learning
• Applied Computing
• intelligent tutoring system
• COVID-19