Application of learning analytics for sequential patterns detection associated with moments of distraction in students in e‐learning platforms

Abstract

Owing to the positive impact of technology on education, in recent years, e‐learning platforms have increased in popularity and implementation. However, different authors agree that, despite the benefits of using these platforms, one of the main problems is the distraction during the interaction between the student and the E‐Learning platform, since the students visit web resources as social networks, which is not related to their learning process. to identify patterns associated with moments of distraction that students can present on these platforms. Following the above, this study aims to apply Learning Analytics using the Density‐based spatial clustering of applications with noise (DBSCAN) clustering technique and the PrefixSpan algorithm, to identify frequent sequential patterns of interaction, before a distraction event of the student from their interaction with a virtual learning platform. From this, clustering with a Silhouette coefficient close to 0.75 is obtained, presenting good separation, high cohesion, and compaction in the generated groups. Additionally, a slight characterization of each group is achieved based on the most frequent sequential patterns of interaction. Finally, in this study, a successful integration between DBSCAN and PrefixSpan is achieved, to identify sequential interaction patterns associated with moments of distraction in E‐Learning platforms, which generates a starting point for later studies to make the sequential recommendation of educational resources from these patterns.