摘要
成绩预测是指通过数据建模的方式来预测学习者未来的学习趋势。但现有研究中大多是基于单一算法构建成绩预测模型,忽略了不同方法的优势与适用领域。因此,基于Stacking集成学习的方法,构建组合模型对大学生的学业成绩进行预测,使学习者及早意识到学业危机。以某高校信息与计算科学专业学生的成绩为研究对象进行大量的随机实验。实验结果表明,在测试样本占比10%~60%的情况下,组合模型的MAE值均小于基于单一算法构建的成绩预测模型。该方法可以从数据空间、数据结构等方面对学生的学习成绩进行全方位多角度分析,提升学生学习动机。
Achievement prediction refers to the method of data modeling to predict the future learning trend of learners.However,most of the existing researches are based on a single algorithm to build a performance prediction model,ignoring the advantages and application fields of different methods.Therefore,based on the method of stacking integrated learning,a combination model is constructed to predict college students’academic performance and make learners aware of academic crisis in advance.A large number of random experiments were conducted based on the scores of students majoring in information and computing science in a university.The experimental results show that the MAE value of the combined model is smaller than that of the grade prediction model based on a single algorithm when the proportion of test samples is 10%~60%.This method can analyze students’academic performance from the aspects of data space and data structure,and improve students’learning motivation.
作者
高艺璐
刘鸿雁
姚明海
李劲松
GAO Yilu;LIU Hongyan;YAO Minghai;LI Jinsong(School of Educational Science,Bohai University,Jinzhou 121013,China;School of Information Science and Technology,Bohai University,Jinzhou 121013,China)
出处
《电子设计工程》
2024年第24期63-67,共5页
Electronic Design Engineering
基金
辽宁省社会科学规划基金项目(L22BTJ002)
辽宁省研究生教育教学改革项目(LNYJG2022405,LNYJG2023285)
渤海大学硕士研究生创新基金项目(YJC2022-027)。
