.png){kind=icon}
.png){kind=icon}
Return
Using learner engagement patterns for predictive modeling in a MicroMasters program
.png){kind=icon}
PRE
AI
DOI：10.1080/01587919.2024.2441252.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This study explores the predictors of learner success in a MicroMasters focusing on behaviors across multiple courses that influence the final comprehensive exam outcome. MicroMasters, a term trademarked by edX, are a series of massive open online courses (MOOCs) that stack into a final credential that can translate into formal academic credit. Using a bi-directional Long Short-Term Memory (LSTM) model, we analyzed engagement data from 1,603 massive open online course (MOOC) learners, achieving 96.6% accuracy in predicting pass/fail outcomes. SHAP (SHapley Additive exPlanations) values identified key behavioral predictors, emphasizing the importance of consistent engagement, active forum participation, and interaction with assessments. Learners who regularly interacted with course materials, attempted questions, and engaged in discussions were more likely to pass the exam and earn the MicroMasters. Performance in foundational courses emerged as a critical predictor of success. These findings address gaps in understanding the temporal and behavioral dimensions of learner engagement and offer actionable strategies for course facilitators to implement targeted interventions that improve completion rates and support student success. This study highlights the potential of machine learning to enhance learner support and inform instructional design, ultimately advancing the quality and scalability of microcredential programs in online, competency-based education pathways.
Keywords:
MOOC
MicroMasters
MOOC learner success
mesocredential
microcredential
MOOC learner engagement
Journal
IF:
3
Papers:
619
Citations:
1.8K
