Diagnostic method improves predicting student learning and skills over time
DIKT: A Diagnostic and Interpretable Knowledge Tracing Framework for Prediction and Model-based Learning Analysis
Machine LearningArtificial Intelligence
Summary
Predicting how well students learn and answering questions helps educational software adapt better, but most methods don’t explain why a prediction is made or how to improve learning. The authors created DIKT, a system that not only predicts student answers but also shows how their skills improve or struggle over time. It breaks down the chance of getting a question right by specific skills and lets teachers or programs try out ‘what-if’ scenarios to see how changes might help learning. Their tests show DIKT predicts better than previous methods and gives useful insights for improving education.
What this means in practice
- •For educational software developers: Integrate DIKT to enhance adaptive learning platforms by providing interpretable skill diagnostics and better student response predictions.
- •For training program designers: Use DIKT’s learning trajectory insights to design interventions that target specific student skill weaknesses over time.
Authors
Yi-Qiang Wua, Shaohui Chen, Hanlin Li, Baolin Yi
Abstract
Knowledge tracing is a key component of intelligent educational systems. However, most existing methods mainly focus on response prediction, providing limited support for explaining predicted outcomes and improving students'learning states. This paper proposes Diagnostic and Interpretable Knowledge Tracing (DIKT), which integrates response prediction, skill-level diagnosis, and model-based learning analysis. DIKT generates bounded cognitive-state trajectories and decomposes the final response probability into skill-wise probabilities. The trajectories reveal students'evolving strengths and weaknesses, while the skill-wise probabilities identify the skills that most strongly constrain the prediction. Based on these diagnostic results, DIKT supports direct cognitive-state intervention and effective-learning-duration-based what-if analysis to examine how hypothetical adjustments affect current predictions and subsequent state evolution. DIKT combines Bayesian multi-head attention, conditional variational recurrent inference, a logistic-growth decoder, and multiplicative skill-wise prediction. Experiments on three benchmark datasets show that DIKT outperforms all evaluated baselines, with an average absolute AUC gain of 2.04 percentage points. Other experiments further demonstrate that DIKT supports educationally meaningful diagnostic analysis and provides informative diagnostic indicators for downstream applications, extending knowledge tracing from response prediction to interpretable diagnosis and model-based learning analysis.