基于特征选择的学生成绩预测方法研究  

Research on student scores prediction method based on feature selection

作　　者：刘晓雲 刘鸿雁 李劲松 王冠帮 LIU Xiao-yun;LIU Hong-yan;LI Jin-song;WANG Guan-bang(School of Education Science,Bohai University,Jinzhou 121013,Liaoning Province,China;School of Information Science and Technology,Bohai University,Jinzhou 121013,Liaoning Province,China)

机构地区：[1]渤海大学研究生学院,辽宁锦州121013 [2]渤海大学信息科学与技术学院,辽宁锦州121013

出　　处：《信息技术》2023年第10期17-22,共6页Information Technology

基　　金：教育部产学合作协同育人项目(202002151003,20210-2085003);辽宁省教育厅科研项目(WJ2020004,LJ2020003);辽宁省教学改革项目(10167-649,10167-636);研究生创新基金项目(YJC2021-030);研究生教育教学改革项目(YJG20210002,YJG20210005)。

摘　　要：学习成绩是反映学习效果和教学质量的重要指标,对成绩进行预测可改进学习和教学方法,进而提高教学质量,如何准确地预测成绩已成为教育数据挖掘领域的一个热点研究课题。为提高预测的准确度,提出了基于特征选择的成绩预测方法。首先利用序列前向选择算法对样本数据进行特征选择,从而选出最优特征子集来构建多元线性回归预测模型,再利用模型对成绩进行预测。为检验方法的有效性,在真实数据集上进行了验证,实验结果表明:文中方法具有更高的预测精度,可以为改进教学方法和教学质量提供数据支持。Academic scores is an important indicator to reflect the learning effect and teaching quality,and the prediction of academic scores can improve the learning and teaching methods,thus improving the teaching quality.How to accurately predict grades has become a hot research topic in the field of educational data mining.In order to improve the accuracy of prediction,a prediction method based on feature selection is proposed.Firstly,the sequence forward selection algorithm is used to select the features of the sample data,and then the optimal feature subset is selected to construct the multiple linear regression prediction model.Finally,the model is used to predict the results.In order to verify the validity of the method,the experiment is carried out,and the results show that the method based on feature selection has higher prediction accuracy and can provide data support for improving teaching methods and teaching quality.

关 键 词：数据挖掘 特征选择 数据降维 多元线性回归 成绩预测 

分 类 号：TP3-05[自动化与计算机技术—计算机科学与技术] G420[文化科学—课程与教学论]