August 15, 2025Open Access

Capturing Session-to-Session Dynamics of Learning and Forgetting: Testing the Limits of Knowledge Tracing Models

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Observational analysis reveals knowledge tracing models struggle with learning and forgetting in students, suggesting a need for improvement.

Selected models capture qualitative trends of learning across sessions, but struggle to predict future behavior effectively.
Bayesian Knowledge Tracing and Additive Factors Model show relatively acceptable fit metrics in retrospective analysis, but fail in prospective predictions.
The models inadequately account for cognitive principles like the spacing effect and patterns of forgetting across sessions.
Enhancing cognitive approaches with computational modeling may lead to better educational outcomes and improved understanding of learning processes.

Schuetze et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

Observational analysis reveals knowledge tracing models struggle with learning and forgetting in students, suggesting a need for improvement.