摘要
为拓展灰色关联分析在面板数据中的应用,针对面板数据的三维特征,将面板数据转化为空间网格.然后基于二维灰色投影关联度,利用空间网格在时间和对象平面上的投影面积,定义了三维灰色投影面积关联度模型.为了解决因为维度不同而导致的关联度差异过大问题,引入灰色聚类检验模型,给出了三维灰色投影面积关联度的面板聚类步骤.最后通过对珠三角四市的空气质量问题的聚类分析,结果表明各类别的差异明显,层次划分清楚,证实了该方法的有效性和实用性.
To expand the application of grey relational analysis in panel data,the threedimensional characteristics of panel data are discussed in this paper,which are converted into spatial grids.Then based on two-dimensional grey projection correlation degree,a threedimensional grey correlation model with projection area is defined by using the projection area of space grid on time and object plane.In order to solve the problem of too large difference in correlation degree caused by different dimensions,by introducing the grey cluster testing model,the panel clustering steps for three-dimensional grey correlation degree with projection area are given.Finally,through cluster analysis of air quality problems in four cities of the Pearl River Delta,the results show that there are exist obvious differences among different categories,and the division of different levels is very clear,which confirm the validity and practicability of the proposed method.
作者
李亚男
LI Ya-nan(Department of Basic Courses,Guangdong Polytechnic College,Zhaoqing 526100,China)
出处
《数学的实践与认识》
北大核心
2020年第14期69-75,共7页
Mathematics in Practice and Theory
基金
广东省教育厅科研项目(2018KQNCX315,2018KTSCX276)
广东理工学院科技项目(GKJ2017025,GKJ2018013)。
关键词
面板数据
灰色投影面积关联度
灰色聚类方法
空气质量
panel data
grey correlation degree with projection area
grey clustering method
air quality
