Scholarlink

<p>Bayesian Knowledge Tracing (BKT) is a popular model in cognitive science and educational technology used to estimate learners’ evolving knowledge states over time, widely used due to its interpretability, simplicity, and theoretical alignment with learning processes. However, existing implementations are primarily available in Python or C++, creating accessibility challenges for researchers working in R–a language extensively used in education and psychology. To address this gap, we present the BKT R package, a comprehensive tool that implements the standard BKT model along with five widely used variants: Prior Per Student (PPS), Item Order Effect (IOE), Item Difficulty Effect (IDE), Item Learning Effect (ILE), and Learning and Forgetting Behavior (LFB). The package features parameter estimation via the Expectation-Maximization algorithm, seamless integration with R’s data manipulation and visualization tools, and support for example datasets and streamlined model output. We validate the package through simulation studies and application to real-world data from the Cognitive Tutor system, demonstrating high fidelity with existing Python implementations and robust parameter recovery across scenarios. This work enhances the accessibility, reproducibility, and flexibility of BKT modeling for R users in behavioral sciences.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

BKT: A Bayesian knowledge tracing package for the R environment

  • 期刊论文

  • 发表日期：

    2026年4月22日

查看全文
  • Yuhao Yuan,
  • Biying Zhou,
  • Jia Qi,
  • Nanyu Luo,
  • Feng Ji

摘要

Bayesian Knowledge Tracing (BKT) is a popular model in cognitive science and educational technology used to estimate learners’ evolving knowledge states over time, widely used due to its interpretability, simplicity, and theoretical alignment with learning processes. However, existing implementations are primarily available in Python or C++, creating accessibility challenges for researchers working in R–a language extensively used in education and psychology. To address this gap, we present the BKT R package, a comprehensive tool that implements the standard BKT model along with five widely used variants: Prior Per Student (PPS), Item Order Effect (IOE), Item Difficulty Effect (IDE), Item Learning Effect (ILE), and Learning and Forgetting Behavior (LFB). The package features parameter estimation via the Expectation-Maximization algorithm, seamless integration with R’s data manipulation and visualization tools, and support for example datasets and streamlined model output. We validate the package through simulation studies and application to real-world data from the Cognitive Tutor system, demonstrating high fidelity with existing Python implementations and robust parameter recovery across scenarios. This work enhances the accessibility, reproducibility, and flexibility of BKT modeling for R users in behavioral sciences.