摘要
校园一卡通系统正被广泛应用于各大高校并产生海量时空数据,利用数据挖掘技术可以从中挖掘出隐藏的大学生在校行为模式。文章首先利用映射时空数据得到学生共现数据;然后采用超几何分布与多重假设检验的方法挖掘大学生社交关系;最后基于交友数量和亲密度得到三种行为模式,发现封闭型行为模式更易形成脆弱的交往关系。根据大学生在校行为模式,学校可以准确掌握学生动态,为其提供更加人性化的管理和精准化的服务。
The campus smart card system is widely used in various universities and generates massive amounts of Spatio-temporal data.Data mining technology can be used to mine hidden behavior patterns.Firstly,the co-occurrence data is obtained by mapping spatio-temporal data;Then,the hypergeometric distribution and multiple hypothesis testing methods are used to mine college students'social relations;Finally,based on the number of friends and intimacy,three behavior patterns are obtained,and it is found that isolated behavior patterns are more likely to form fragile social relations.According to the behavior patterns,the dynamics of student can be accurately grasped,and more humane management and precise services can be provided.
作者
刘涛
LIU Tao(School of Computer and Big Data Science,Jiujiang University,Jiujiang,Jiangxi 332005)
出处
《科教导刊》
2023年第18期143-145,共3页
The Guide Of Science & Education
基金
江西省教育科学“十三五”规划2020年度课题(20YB206)
关键词
时空数据
超几何分布
行为模式
spatio-temporal data
hypergeometric distribution
behavior patterns
