Automated Knowledge Component Generation for Interpretable Knowledge Tracing in Coding Problems

This paper develops an automated LLM-based pipeline (KCGen-KT) to generate knowledge components for programming problems and uses them for knowledge tracing to predict student mastery levels and future performance. The system is evaluated on real-world student code submission datasets, demonstrating improved prediction accuracy over human-written KCs and existing knowledge tracing methods.

Knowledge components (KCs) mapped to problems help model student learning, tracking their mastery levels on fine-grained skills thereby facilitating personalized learning and feedback in online learning platforms. However, crafting and tagging KCs to problems, traditionally performed by human domain experts, is highly labor intensive. We present an automated, LLM-based pipeline for KC generation and tagging for open-ended programming problems. We also develop an LLM-based knowledge tracing (KT)

Study Type

Source

Framework Categories

Tool Types

Tags