Released learner performance prediction code and public dataset links¶

Element · All elements
Evidence · no claims cited

Description¶

The article releases its experimental code and links to the nine public benchmark datasets used in the comparison. The implementations include deep learning algorithms in PyTorch, the logistic regression implementation from scikit-learn, and the BKT+ implementation from Khajah et al. 2016 in C++; the authors state "Our code and links to public datasets we used are freely available on GitHub". This artifact supports replication of the cross-dataset comparison and ablation studies.

Design Implications¶

Context¶

Requirements¶

• Deep learning algorithms require the PyTorch implementation; logistic regression uses scikit-learn; BKT+ requires the C++ implementation from Khajah et al. 2016

Constraints¶

• Algorithms requiring timestamps (like DAS3H) were only applicable to the subset of datasets containing timestamp information

Target Learners¶

• educational data mining researchers and learning engineers replicating learner performance model comparisons

Target Learning Goals¶

• accurate prediction of learner correctness in intelligent tutoring systems

Claims¶

Related Elements¶

• EduData and EduKTM Open-Source Knowledge Tracing Libraries

Examples¶

-

Key Sources¶

• Theophile Gervet, Ken Koedinger, Jeff Schneider, Tom Mitchell (2020). When is Deep Learning the Best Approach to Knowledge Tracing? Journal of Educational Data Mining, Volume 12, No 3. https://github.com/theophilee/learner-performance-prediction