| 摘 要: 针对教育大数据应用中异构数据源分散、教学评价体系单一等问题,构建了基于多源大数据与机器学
习的教学评价系统。系统对学生行为日志、课程元数据、成绩和教师评价等数据进行跨源数据融合,采用了梯度提
升树回归、集成学习等机器学习方法(知识掌握度预测模块R2=0.993,成绩与时间预测模块R2>0.87)。消融实验
验证并解释了任务完成率(0.35)和学习效率(0.25)是关键因素。实验结果表明,该系统为教育决策提供了数据驱动
依据,有助于实现科学化教学。 |
| 关键词: 教育大数据 学习分析 个性化推荐 教师绩效 机器学习 |
|
中图分类号:
文献标识码: A
|
| 基金项目: 中国教育技术协会开放与远程教育专业委员会2025年度科研课题(KYGZYB25001) |
|
| Research and Implementation of Teaching Evaluation System Based on Multi-source Big Data and Machine Learning |
|
ZHANG Yujie
|
(Beijing National Accounting Institute, Beijing 101312, China)
zhangyj@mail.nai.edu.cn
|
| Abstract: To address the issues of scattered heterogeneous data sources and unitary teaching evaluation systems
in educational big data applications, this study constructs a mult-i source big data and machine learning-based teaching
evaluation system. The system performs cross-source data fusion on student behavior logs, course metadata, grades, and
teacher evaluation data. Machine learning methods including gradient boosting tree regression and ensemble learning
are employed (knowledge mastery prediction module R
2 = 0.993, grade and time prediction modules R
2 > 0.87).
Ablation experiments validate and explain that task completion rate (0.35) and learning efficiency (0.25) are key factors.
Experimental results demonstrate that the system provides data-driven evidence for educational decision-making,
contributing to the realization of scientific teaching. |
| Keywords: educational big data learning analytics personalized recommendation teacher performance machine learning |
