Hello EDM Community,
We are inviting you to participate in this year’s EDM Cup.
Understanding how students’ behaviors affect their long-term performance has a wide range of applications. Knowledge tracing, item response theory, affect detection, dropout prediction, gaming detection, and many other common methods explored in educational data mining stem from the idea that the actions students take while they complete their assignments can be used to model latent variables that predict learning and performance.
In this competition, you are tasked with predicting students’ scores on end-of-unit assignment problems given their click-stream data across all the in-unit assignments they completed previously. This task will rely heavily on the methods used to extract relevant features from millions of actions taken by students within the ASSISTments online learning platform as they completed their mathematics assignments. Additionally, this data set includes information on the curricula, assignments, problems, and tutoring provided to the students to help inform your predictions.
Participants in this competition are invited to submit to a special issue of The Journal of Educational Data Mining. Papers should focus on the approach used in this competition to understand and model students’ behavior, with emphasis on interpretable insights and how these insights can help individual students and improve online learning in general. Acceptance to the special issue will be based on a combination of paper quality and leaderboard rank.
The competition, hosted on Kaggle, can be found here.
Submissions to the special issue are managed through EasyChair and can be submitted here through the EDM2023 EDM Cup Track.
The competition ends May 31st, 2023 23:59:00 UTC.
Submissions to the special issue are due June 15th, 2023 AOE.