摘要
为提升学生成绩评估水平和就业成功率,提出了学生成绩就业集成预测系统.该系统分为数据层、应用层、分析层,协调实现学生数据管理、统计分析和管理服务.首先,选取9个学生就业预测指标和相应的历史学习数据.其次,基于累积自回归滑动平均模型对学生数据进行处理,将非平稳时间序列数据利用微分变换转换为平滑的时间序列.再次,提出基于LSSVM集成模型对样本进行训练.最后,基于550个学生数据,对所提模型进行验证.实验结果表明,所提LSSVM集成模型预测准确率为到92.7%,较RF、MLR、ANN和SVM相比分别提升13.5%、6.4%、17.3%、2.8%.
In order to improve students'achievement evaluation level and employment success rate,this paper proposes an integrated prediction system for students'achievement and employment.The system is divided into data layer,application layer and analysis layer to coordinate and realize student data management,statistical analysis and management services.Firstly,nine student employment prediction indicators and corresponding historical learning data are selected.Secondly,the student data are processed based on the cumulative autoregressive moving average model,and the non-stationary time series data are transformed into smooth time series by differential transformation.Thirdly,an integrated model based on LSSVM is proposed to train the samples.Finally,based on 550 student data,the proposed model is verified.The experimental results show that the prediction accuracy of the proposed LSSVM integrated model is 92.7%,which is 13.5%,6.4%,17.3%and 2.8%higher than that of RF,MLR,Ann and SVM respectively.
作者
曲径
QU Jing(School of Economics and management,Northeast Petroleum University,Daqing 163318,China)
出处
《云南民族大学学报（自然科学版）》
CAS
2023年第3期382-387,共6页
Journal of Yunnan Minzu University:Natural Sciences Edition
基金
中国高等教育学会高校辅导员队伍建设与发展研究专项课题(22FD0218)
黑龙江省教育科学规划重点课题(GJB13200402).
关键词
教育数据挖掘
自回归滑动平均
最小二乘支持向量机
集成模型
educational data mining
autoregressive moving average
least squares support vector machine
integrated model
