基于Canopy-K-means算法的高校贫困生预测的研究  

Research on Prediction of Poor College Students Based on Canopy-K-means Algorithms

作　　者：王全民[1] 张书军 WANG Quanmin;ZHANG Shujun(School of Computer,Beijing University of Technology,Beijing 100124)

机构地区：[1]北京工业大学计算机学院,北京100124

出　　处：《计算机与数字工程》2020年第12期3012-3016,3041,共6页Computer & Digital Engineering

摘　　要：高校贫困生的认定和和资助工作对于高等人才的培养、减轻贫困生家庭经济负担是非常重要的;如何做到精准扶贫一直是高校贫苦生认定和资助工作的重点和难点。基于校园一卡通学生消费数据、上网日志等多模态数据,提出了一种基于Canopy-K-means算法的高校贫困生预测的方法。该方法通过引入Canopy改进的聚类算法,得到贫困生所属类别,并将该类学生与实际的贫困生作对比,分析贫困生在校消费习惯和上网行为。该实验能有效地对贫困生进行分类,为学校贫困生认定提供辅助决策作用。The identification and funding of poor students in colleges and universities is very important for the cultivation of higher talents and the relief of the economic burden of poor families.How to achieve accurate poverty alleviation has always been the focus and difficulty of the identification and funding of poor students in colleges and universities.Based on multi-modal data such as campus card student consumption data and online log,a method based on Canopy-K-means algorithm proposed for predicting poor students in colleges and universities.By introducing Canopy's improved clustering algorithm,this method obtains the categories of poor students,and compares these students with actual poverty-stricken students,and analyzes the consumption habits and online behaviors of poor students.This experiment can effectively classify poor students and provide assistant decision-making for the school's poor students.

关 键 词：贫困生认定 多模态数据 数据挖掘 Canopy-K-means算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]