摘要
随着智能学业测评的广泛应用,其在教学反馈和数据处理方面展现出明显优势,但背后的建模与预测机制仍面临诸多挑战,亟需进一步优化和拓展。为此,文章首先系统地回顾了智能学业测评与学习表现预测的理论基础与技术演进,据此界定研究目标与方法论框架;然后,文章以552名学生长达三个月的练习数据与324名学生的考试数据为样本,经数据处理后筛选并组合与最终考核成绩关系较密切的预测变量;接下来,文章构建逻辑回归、决策树、神经网络、支持向量机四类预测模型,并通过准确率、精确度、AUC等指标评估其性能;最后,文章采用装袋、提升及随机森林三种集成策略进行模型优化,发现集成学习显著提升了预测性能,其中Boosting对神经网络的增益最大,精确率提升11.2个百分点。结果表明,文章所构建的在线学习表现预测模型兼具高预测准确率、可解释性与易操作等特点,可为精准化学习诊断与即时教学干预提供理论支撑和实践路径。
The widespread application of intelligent academic assessment has demonstrated significant advantages in teaching feedback and data processing,yet the underlying modeling and prediction mechanisms still face numerous challenges that urgently need further optimization and expansion.To address this,the paper first systematically reviewed the theoretical foundations and technological trajectories of intelligent academic assessment and learning performance prediction,and accordingly defined the research objectives and methodological framework.Then,the paper took the practice data of 552 students over a three-month period and the examination data of 324 students as samples to identify and combine the most relevant predictive variables for final assessment scores.Subsequently,the four types of prediction models including logistic regression,decision tree,neural network,and support vector machine were constructed and evaluated based on the indicators of accuracy rate,precision degree,and AUC.Finally,three ensemble strategies,such as Bagging,Boosting and Random Forest,were applied to optimize the models.It was found that ensemble learning significantly improved the predictive performance,with Boosting yielding the largest enhancement for the neural network,as reflected by an 11.2%increase in precision.The findings demonstrated the constructed intelligent prediction model exhibited high prediction accuracy,interpretability and ease of operation,which can provide theoretical support and practical pathways for precise learning diagnostics and timely instructional interventions.
作者
冷静
吴子豪
都洋岚
卢弘焕
LENG Jing;WU Zi-Hao;DU Yang-Lan;LU Hong-Huan(Faculty of Education,East China Normal University,Shanghai,China 200062;High School Affiliated to University of Shanghai for Science and Technology,Shanghai,China 200093;School of Economics and Management,East China Normal University,Shanghai,China 200062)
出处
《现代教育技术》
2025年第8期87-96,共10页
Modern Educational Technology
基金
中央高校基本科研业务费项目华东师范大学哲学社会科学创新团队项目“智能化读写评价技术在语文教育中的运用研究”(项目编号:2024QKT004)
中央高校基本业务费项目华东师范大学青年预研究项目“基于多模态大模型的批判性思维测评研究”(项目编号:2024ECNU-YYJ026)的阶段性研究成果。
关键词
智能学业测评
机器学习
集成学习
在线表现预测模型
虚拟仿真系统
intelligent academic assessment
machine learning
ensemble learning
online performance prediction model
virtual simulation system
