Investigating the impact of emotions on learners’ success and dropout prediction in a MOOC

Massive Open Online Courses (MOOCs) have transformed education by providing accessible and flexible learning opportunities. These platforms generate substantial data through learners’ interactions with multimedia resources. Data mining techniques are applied to analyse this data for various purposes, such as predicting dropout risks and learner success. However, there is a notable gap in current research regarding the incorporation of emotional data from learners during e-learning. This study addresses this gap by investigating the influence of emotional data on learner behaviour within MOOC platforms. We employ preprocessing techniques and classical machine learning algorithms, specifically Multilayer Perceptron (MLP) and Bidirectional Long Short-Term Memory (BILSTM), on data extracted from the French Project Management MOOC (GDP). Our aim is to predict learners’ dropout and success using navigation data and emotional features. The findings demonstrate that navigation and emotional data can independently and effectively predict learner dropout and success. Interestingly, integrating emotional features with navigation data does not significantly improve predictive performance compared to using either data source individually. This outcome suggests that emotional patterns may already be implicitly embedded within navigation behaviours.