基于异质Stacking集成学习的大学生学业风险预测及早预警  

Academic Risk Prediction for College Students Based on Stacking Integrated Learning and Early Warning

作　　者：杨舟 崔彩霞[1] YANG Zhou;CUI Caixia(School of Computer Science and Technology,Taiyuan Normal University,Jinzhong Shanxi 030619,China)

机构地区：[1]太原师范学院计算机科学与技术学院,山西晋中030619

出　　处：《太原师范学院学报(自然科学版)》2025年第1期22-29,39,共9页Journal of Taiyuan Normal University:Natural Science Edition

摘　　要：针对学生学业风险预测提出一种基于异质Stacking集成学习模型,对有潜在学业危机的学生进行早预警,减少学生的辍学率和教育资源的浪费,进行有效的补救及调整.该模型运用皮尔逊相关分析技术对原始数据集进行特征选择,然后通过支持向量机、极端梯度提升树、K近邻和随机森林等4个基础模型来构建初级分类器,逻辑回归为次级分类器两层结构来构建Stacking集成学习模型.实验结果表明,与传统机器学习模型相比,该模型在准确率、精确率、召回率、F1分数和G-Mean等指标上有明显提升.A heterogeneous Stacking ensemble learning model is proposed for predicting students'academic risks,which can provide early warning for students with potential academic crisis,reduce the dropout rate and waste of educational resources,and make effective remediation and adjustments.The model uses Pearson correlation analysis technology to select features of the original data set,and then constructs a primary classifier through four basic models such as support vector machine,extreme gradient boosting tree,K nearest neighbor and random forest,and uses logistic regression as a secondary classifier to construct a stacking ensemble learning model.The experimental results show that compared with the traditional machine learning model,the model has significant improvements in accuracy,precision,recall,F1 score and G-Mean.

关 键 词：STACKING 集成学习 学业风险预测 早预警 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]