基于梯度回归模型的学生成绩预测  

Students′Grade Prediction Based on Gradient Regression Model

作　　者：贾丙静 孙瑜 黄心依 赵海燕 段汉根 Jia Bingjing;Sun Yu;Huang Xinyi;Zhao Haiyan;Duan Hangen(College of Information and Network Engineering,Anhui Science and Technology University,Bengbu,Anhui 233000,China;School of Electrical and Information Engineering,Heilongjiang University of Technology,Jixi,Heilongjiang 158100,China)

机构地区：[1]安徽科技学院信息与网络工程学院,安徽蚌埠233000 [2]黑龙江工业学院电气与信息工程学院,黑龙江鸡西158100

出　　处：《黑龙江工业学院学报(综合版)》2025年第2期102-106,共5页Journal of Heilongjiang University of Technology(Comprehensive Edition)

基　　金：安徽省高等学校省级质量工程项目“Java面向对象程序设计”(项目编号:2022xxkc024);安徽科技学院校级教学改革研究项目“知识图谱驱动的计算机程序设计类课程个性化教学研究”(项目编号:Xj2023138)。

摘　　要：大学生在校期间会产生各种纷繁复杂的数据,难以靠单一数据预测学生学习状况。因此,首先构建更为立体的学生行为指标,包括每节课专注时长、每周去图书馆次数、每天睡眠时长、学习目标、上课压力和课程成绩,多角度刻画学生画像,分析对课程成绩影响比较大的特征;然后在梯度回归模型的基础上,预测学生成绩,并在决定系数、平均绝对误差和均方根误差的基础对比各模型的预测结果,实验结果表明,我们所使用的梯度回归模型在各个评价指标上都取得了最好的结果,预测结果可以帮助老师制定个性化的教学方案,从而提升教学质量。College students generate various complex data during their study in school,making it difficult to predict their learning status based on single data.Therefore,this study firstly constructs a three-dimensional student behavior index,including the duration of focus in each class,the number of times to go to the library per week,the length of sleep per day,learning goals,class pressure,and course grades,to depict student profiles from multiple perspectives and analyze the characteristics that have a significant impact on course grades;then,based on the gradient regression model,student grades are predicted,and the prediction results of each model are compared based on the coefficient of determination,mean absolute error,and root mean square error.The experimental results show that the gradient regression model we used achieve the best results in various evaluation indicators,and the prediction results can help teachers improve teaching quality.

关 键 词：JAVA面向对象程序设计 学习行为 梯度回归模型 成绩预测 

分 类 号：TP311.1[自动化与计算机技术—计算机软件与理论] G642[自动化与计算机技术—计算机科学与技术]