The TRACE transformer architecture jointly predicts student course enrollments and grades by encoding course concurrency within semesters. Using ten years of institutional data, the model improves prediction quality by capturing the impact of simultaneous course loads.
HOW THIS AFFECTS YOU
●
researcherThis approach demonstrates the utility of joint prediction tasks over simple sequential modeling in educational datasets.