基于BKT模型的网络教学跟踪评价研究

Research on Tracing Evaluation of Web-Based Teaching Based on BKT Model

当前网络教学中存在教学过程评价难以量化、课程教学进度安排缺少数据支持等问题。随着大数据技术的发展,网络学习行为分析已经取得较大进展,但学习内容的跟踪与评价还比较缺乏。将知识跟踪嵌入网络课程,及时跟踪学生知识掌握情况,将有助于教师发现学生学习问题,调整教学策略;同时也可引导学生将学习的关注点聚焦在知识内容的理解上,而不是分数上。贝叶斯知识跟踪(BKT)模型是一种以知识点为核心构建学生知识模型的方法,具有简捷、预测准确、易于解释的特点。基于BKT公式改进的网络教学跟踪评价模型,可用以课时估算和学习成绩预测。实证分析数据显示,该模型的预测准确率和精确度较高。在实际应用中,BKT知识跟踪功能可单独开发应用,也可与教学平台集成使用,亦可支持线下教学。

In the current web-based teaching,there are problems such as the difficulty in quantifying the evaluation of teaching process and the lack of data support for the teaching schedule.With the development of big data technology,web-based learning behavior analysis has made great progress,but the tracing and evaluation of learning content is still relatively lacking.Embedding knowledge tracing into web-based courses and tracing students’knowledge mastery in time will help teachers discover students’learning problems and adjust teaching strategies.At the same time,it can also guide students to focus on the understanding of knowledge content rather than on scores.Bayesian Knowledge Tracing(BKT)model is a method to construct students’knowledge model with knowledge points as the core.It is simple,accurate and easy to interpret.Based on the improved BKT formula,the web-based teaching tracing evaluation model can be used to estimate course hours and predict learning performances.Empirical analysis data shows that the prediction accuracy and precision of the model are relatively high.In practical applications,BKT knowledge tracing function can be developed and applied separately or integrated with web-based teaching platform,and can also support offline teaching.