基于高校学生行为数据的成绩预警聚类分析

Clustering Analysis of Achievement Early Warning Based on Behavior Data of College Students

随着校园信息化建设的不断深入,校园内各应用系统逐渐增多,运用数据挖掘技术可以从海量的数据中提取潜在有用的信息用于分析高校学生的日常行为与成绩方面的相关性。对一卡通系统、图书馆管理系统、教务系统等多维数据进行采集,基于密度聚类算法,在初始聚类中心选择的基础上结合了距离的度量,重新定义核心点、孤立点、边界点等概念,构建一个改进的密度聚类算法进行数据挖掘分析,达到对学生学习成绩的预警,避免出现挂科、留级等严重的学业问题。

With the continuous deepening of campus information construction,the application systems in the campus are gradually increasing.The data mining technology can be used to extract potential and useful information from massive data to analyze the correlation of college students'daily behavior and achievement.This paper collects the data from one card system,library management system,educational system and other multidimensional data,based on the density clustering algorithm,combines the measurement of distance on the basis of the initial clustering center selection,redefines the concepts of core point,isolated point,boundary point and so on,and builds a improved density clustering algorithm for data mining analysis,so as to achieve early warning of students'academic achievement and avoid serious academic problems such as fail and retardation.