基于K-medoids算法的学生成绩聚类研究  

A Study on Student Performance Evaluation Based on Fast K-medoids Algorithm

作　　者：段桂芹[1] 邹臣嵩[2] DUAN Guiqin;ZOU Chensong(Department of Computer Science,Guangdong Songshan Polytechnic College,Shaoguan 512126;Department of Electrical Engineering,Guangdong Songshan Polytechnic College,Shaoguan 512126)

机构地区：[1]广东松山职业技术学院计算机系,韶关512126 [2]广东松山职业技术学院电气工程系,韶关512126

出　　处：《微型电脑应用》2020年第2期64-66,共3页Microcomputer Applications

基　　金：韶关市科技计划项目(2017CX/K055);广东松山职业技术学院重点科技项目(2018KJZD001).

摘　　要：针对课程之间难度差异大而带来的学生成绩评价"一刀切"的问题,提出了一种基于K-medoids算法的学生成绩评价方法。首先,使用Z-Score对样本集进行标准化,然后,采用K-medoids算法对学生的多门课程成绩进行聚类,最后,使用内部评价指标Sil对多组聚类结果进行评价,得出最优聚类数和最优聚类划分。通过对某高校学生成绩的分析结果表明,该方法能够有效地对学生成绩进行聚类,有助于教学领域向个性化、信息化的方向发展。According to the problem of"one-size-fits-all"student performance evaluation caused by the difficulty difference between courses,a K-medoids algorithm based on student performance evaluation is proposed.Firstly,Z-score is used to standardize the sample set;secondly,K-medoids algorithm is used to cluster students’scores of multiple courses.Finally,the results of multi-group clustering are evaluated by using the internal evaluation index Sil,and the optimal clustering number and optimal clustering are obtained.The result of the analysis of the students’achievement in a university shows that this method can effectively cluster the students’achievement,which is helpful for the development of individualized and information-based teaching field.

关 键 词：成绩评价 聚类分析 K-medoids 聚类评价 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]