基于KPCA的教育特征群集挖掘与分析  

Mining and analysis of educational feature clusters based on KPCA

作　　者：吴宇航 陈凤凤[1] 江玲 万磊 Wu Yuhang;Chen Fengfeng;Jiang Ling;Wan Lei(Gongqing College,Nanchang University,Jiujiang 332020,China)

机构地区：[1]南昌大学共青学院,江西九江332020

出　　处：《无线互联科技》2023年第23期131-135,共5页Wireless Internet Technology

基　　金：江西省大学生创新创业训练计划项目,项目名称:基于KPCA的教育特征群集挖掘与分析,项目编号:S202313430002;江西省教育厅科学技术研究项目,项目名称:互联网+时代基于数据挖掘的混合教育模式及评价体系研究,项目编号:GJJ2203804。

摘　　要：文章针对混合教学数据的特征分析与分级问题,提出了一种基于核主成分分析(Kernel Principal Component Analysis,KPCA)的方法。文章融合传统教学数据和在线教育数据,形成完整的教育特征数据链;利用复杂网络的研究方法和理论模型,构建教育特征数据网络模型;采用非线性KPCA特征抽取算法提取出影响教育质量的群集特征;建立针对教育质量、学生级别进行综合评价的评价模型,并对学生进行分级,从而更好地提供个性化教育干预,提高教学质量。通过实验验证了该方法的有效性和优越性,表明该方法能够有效地提取混合教学数据的非线性特征,并能够实现混合教学数据的分级。This article proposes a method based on Kernel Principal Component Analysis(KPCA)to analyze and classify the characteristics of mixed teaching data.The article combines traditional teaching data and online education data to form a complete chain of educational feature data.It uses research methods and theoretical models of complex networks to construct an educational feature data network model.It adopts a non-linear KPCA feature extraction algorithm to extract cluster features that impact educational quality.It establishes an evaluation model for comprehensive evaluation of educational quality and student levels,and classifies students in order to provide personalized educational interventions and improve teaching quality.The effectiveness and superiority of this method are verified through experiments,demonstrating its ability to effectively extract non-linear features of mixed teaching data and achieve classification of mixed teaching data.

关 键 词：混合教学 核主成分分析 复杂网络 群集特征 综合评价 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]