| 摘 要: 当前,线上线下的学习过程中,学生的学习活动产生了多种类型的学习数据,基于这些数据,采用机器学习技术能够训练学生成绩的预测模型。文章首先介绍基于机器学习技术构建学生成绩预测模型的基本流程;其次系统地梳理基于机器学习技术的成绩预测工作中使用的学习数据特征,并介绍基于非过程性数据特征和过程性数据特征的成绩预测研究工作;最后分析和总结使用两类学习特征数据进行成绩预测的不足,并对未来研究工作做展望。 |
| 关键词: 非过程性特征;过程性特征;教育数据挖掘;成绩预测 |
|
中图分类号: TP391
文献标识码: A
|
| 基金项目: 山东科技大学青年教师教学拔尖人才培养项目(BJ20200505);山东省教育科学“十四五”规划课题(2021YB028);山东科技大学优秀教学团队建设计划资助项目(JXTD20180503) |
|
| Overview of Learning Data Features Applicable to Student Performance Prediction |
|
ZHANG Feng, CHEN Jingjing
|
(School of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao 266590, China)
zhangfeng@sdust.edu.cn; cjj202083060002@163.com
|
| Abstract: Currently, in the process of online and offline learning, students' learning activities generate various types of learning data. Based on these data, machine learning technology can be used to train prediction models for students' grades. Firstly, the basic process of building student performance prediction model based on machine learning technique is introduced. Then, the learning data features used in performance prediction based on machine learning technique is systematically reviewed, and the research work of performance prediction based on non-process data features and process data features is introduced. Finally, the shortcomings of using the two types of learning data features to predict performance are summarized, and the prospects for future research work is presented. |
| Keywords: non-process features; process features; educational data mining; performance prediction |
