Four large-scale real-world sequential knowledge tracing benchmark datasets used to evaluate Adaptive G-UKT¶

Element · All elements
Evidence · no claims cited

Description¶

The evaluation corpus comprises four large-scale real-world sequential datasets: "ASSISTments2009, Bridge2Algebra2006, Algebra2005, and NIPS34". These datasets supply the historical observation streams of exercise-response pairs over which the model predicts next-response probability and against which state-of-the-art baselines are compared, with the authors reporting particular advantages in sparse data regimes.

Design Implications¶

Context¶

Requirements¶

• Datasets must provide sequential exercise-response interaction streams with item-concept associations

Constraints¶

Target Learners¶

• students whose response logs populate the ASSISTments, Bridge to Algebra, Algebra, and NIPS34 datasets

Target Learning Goals¶

• predicting correctness of future exercise responses from interaction history

Claims¶

Related Elements¶

• KDD Cup 2010 Bridge to Algebra dataset and the hmmsclbl fitting tool
• EduData and EduKTM Open-Source Knowledge Tracing Libraries

Examples¶

-

Key Sources¶

• Jia Nan, Su Weitao, Xian Junrui, Zou Shijia, Xia Yixue. (2026). Adaptive G-UKT: a unified probabilistic framework for knowledge tracing via adaptive graph topology learning and uncertainty-aware Gaussian embeddings. Scientific Reports. https://doi.org/10.1038/s41598-026-50711-y