基于全局中心聚类算法的学生成绩评价研究  

Student performance evaluation based on global center clustering algorithm

作　　者：段桂芹[1] DUAN Guiqin(Depart.of Computer Science,Guangdong Songshan Polytechnic,Shaoguan Guangdong 512126,China)

机构地区：[1]广东松山职业技术学院计算机系,广东韶关512126

出　　处：《智能计算机与应用》2019年第1期80-83,共4页Intelligent Computer and 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：针对课程难度差异大而带来的学生成绩评价难的问题,提出了一种基于全局中心聚类算法的学生成绩评价方法。首先,使用min-max归一化方法对样本集进行预处理;然后,采用全局中心聚类算法对学生的多科成绩进行聚类;最后,使用内部评价指标CH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聚类分析结果表明:该方法能够有效地挖掘出学生多科成绩的分布情况,可为个性化教学的实施提供一种新的思路。In view of the difficulty of students’performance evaluation caused by the difference of curriculum difficulty,a new method of student performance evaluation based on global center clustering algorithm is proposed.Firstly,the sample set is pre-processed by min-max normalization method.Then,the global center algorithm is used to cluster the students’scores of many courses.Finally,the internal evaluation index CH is used to evaluate the results of multi-group clustering,and the optimal clustering number and optimal clustering partition are obtained.Through the cluster analysis of students’performance in a university,the results show that the method can effectively mine the distribution of students’scores in many courses.It can provide a new idea for the implementation of personalized teaching.

关 键 词：成绩评价 全局中心聚类 内部评价指标 数据标准化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]