基于聚类算法的学生学籍数据分析  

Analysis of Student Status Data Based on Clustering Algorithm

作　　者：罗晓军 LUO Xiaojun(Liming Vocational University,Quanzhou,Fujian,China 362000)

机构地区：[1]黎明职业大学,福建泉州362000

出　　处：《湖南邮电职业技术学院学报》2023年第4期52-56,共5页Journal of Hunan Post and Telecommunication College

基　　金：2021年度黎明职业大学校级一般课题“数据挖掘技术在高职院校招生决策中的应用”(课题编号:LT202118);2022年度福建省中青年教师科研项目(高校教育信息化专项)“‘互联网+’教学背景下高校学业预警研究”(项目编号:JAT220843)。

摘　　要：为提升高职院校的生源质量和学生的学业成绩,更好地培养人才和服务地方经济产业发展,通过DBSCAN算法聚类分析学籍数据,计算相似度时使用曼哈顿距离代替欧式距离运算,避免高维数据运算时可能产生的维数灾难,最后以某高职院校2019到2022届毕业生的学籍数据作为实验分析对象,预测绩点所归属的簇,在聚类生成的簇中挖掘出优秀学生共性特征,以及生源数据与学业数据之间的关系。In order to improve the quality of students and students'academic performance in higher vocational colleges,better cultivate talents and serve local economic and industrial development,DBSCAN algorithm is used for clustering analysis of student status data,and Manhattan distance is used instead of Euclidean distance operation when calculating similarity,so as to avoid possible curse of dimensionality during high-dimensional data operation.Finally,taking the student status data of a higher vocational college graduates from 2019 to 2022 as the experimental analysis object,predicting the cluster to GPA belong,the common characteristics of excellent students are excavated from clusters generated by clustering,and the relationship between student source data and academic data is also excavated.

关 键 词：生源质量 人才培养 数据挖掘 DBSCAN聚类 

分 类 号：TP311.12[自动化与计算机技术—计算机软件与理论]