基于K-means聚类算法的学生表现数据分析及预测建模研究

Modeling of Student Performance Data Based on K-means Clustering Algorithm

通过对学生生活、学习、活动等行为特征数据分析挖掘,采用改良的K-means聚类算法建立学生表现类别模型,实现根据学生表现数据将学生进行分类。选择学生“德育成绩、体育成绩、智育成绩、竞赛等级、贫困生等级、奖学金等级”6个属性数据作为特征评价指标。针对高校学生管理系统类别放多造成的数据重复、缺失、存储类型不一致等问题,对数据清洗、集成和变换数据存储格式,得到满足K-means算法的输入数据。

This paper uses the improved K-means clustering algorithm to establish the student performance category model by analyzing and mining the behavior characteristic data of students’life,study and activity,and realizes the classification of students’performance data.The six attribute data of“moral education achievement,sports achievement,intellectual education achievement,competition grade,poor student grade,scholarship grade”are selected as the characteristic evaluation indexes.Aiming at the problems of data duplication,loss and inconsistency of storage types caused by the overpopulation of student management system in colleges and universities,the data cleaning,integration and transformation of data storage format are completed to satisfy the requirement of input data in the K-means algorithm.