Résumé

During the COVID-19 pandemic, most people around the world started using internet and online technologies for multiple purposes. A very common digital transformation is online learning or e-learning. Despite its importance, online learning has several issues compared to standard learning. One of the most issues of concern is cheating between students during online exams. This paper aims to address cheating problem by proposing a novel method called WeCheat. It automates the cheating detection process on e-learning platforms, such as Google Forms. The target for the detection of cheating-groups are online exams. To the best of our knowledge, this is the first time someone proposes such a system that detects cheating, particularly during the COVID-19 pandemic. To comply with the set goal, a clustering-based solution is proposed. This Method employs MeanShift, a non-parametric and density-based clustering algorithm for this task. Validation of this approach is evidenced by the application which performs excellently on an arbitrary number of features in the cheating detection problem in terms of accuracy. This application helps educational institutions to address cheating cases and at the same time offers the opportunity to focus more on the use of anti-cheating logistics.

Détails

Actions