摘要
面向高校多源异构教学数据,构建了以标签体系为核心的学生画像系统。系统设计涵盖数据接入与标准化结构建模,采用监督学习算法完成特征提取与多标签分类模型训练,形成包含认知能力、学习行为等维度的结构化学生画像。通过将随机森林(Random Forest,RF)+支持向量机(Support Vector Machine,SVM)+XGBoost组合模型与其他单一算法对比可知,系统在预测精度、F_1值、响应效率等方面表现优异,特别适用于个性化教学分析与动态干预。
For multi-source heterogeneous teaching data in higher education institutions,a student profiling system centered on a tagging system has been constructed.The system design encompasses data integration and standardized structural modeling,utilizing supervised learning algorithms for feature extraction and multi-label classification model training,resulting in a structured student profiling system that includes dimensions such as cognitive abilities and learning behaviors.Through experimental comparisons between the Random Forest(RF)+Support Vector Machine(SVM)+XGBoost combined model and other single algorithms,the system demonstrates superior performance in terms of prediction accuracy,F1 score,and response efficiency,making it particularly suitable for personalized teaching analysis and dynamic intervention.
作者
赵建伟
ZHAO Jianwei(Liaocheng Vocational and Technical College,Liaocheng,Shandong 252024,China)
出处
《智能物联技术》
2025年第6期91-95,共5页
Technology of Io T& AI
关键词
学生画像
机器学习
系统设计
特征工程
模型评估
student portrait
machine learning
system design
feature engineering
evaluation
