首页|函数型数据视角下基于伪分位数聚类的空气质量治理区域划分

函数型数据视角下基于伪分位数聚类的空气质量治理区域划分

扫码查看
近几年,空气污染与质量问题受到广泛关注.由于我国各大城市空气质量差异较大,区域性特征明显,所以划分空气质量区域,实施针对性空气质量防治措施,对改善和提高我国各地区空气质量具有重要的现实意义.本文以我国312个地级市2015年至2019年空气质量AQI逐日数据,使用伪分位数聚类,即Expectlie曲线聚类和M分位数函数型数据聚类方法,对各地级市空气质量进行研究和治理区域划分.根据两种聚类结果最终划分为9个不同的空气质量治理区域,并针对各区域特点提出污染防治措施.
Regional Division of Air Quality Governance Based on Pseudo-Quantile Clustering with the View of Functional Data
In recent years,air pollution and quality problems have been widely concerned.Due to the large difference of air quality in China's major cities and obvious regional characteristics,it is of great practical significance to divide air quality regions and implement targeted air quality prevention and control measures to improve air quality in China's various regions.In this paper,based on the daily AIR quality AQI data of 312 prefecture-level cities from 2015 to 2019,pseudo-quantile clustering(Expectlie curve clustering and M-quantile functional data clustering)was used to study the air quality of each prefecture-level city and divide the air quality control regions.According to the two kinds of clustering results,9 different air quality control regions were finally divided,and pollution prevention measures were proposed according to the regional characteristics of each region.

expectile curvesM-quantilepseudo-quantilefunctional data clusteringair quality

梁永玉、曹苏周、周梦雨、田茂再

展开 >

临夏县统计局,甘肃临夏 731800

西安思源学院基础部,陕西西安 710038

新疆财经大学统计与数据科学学院,新疆乌鲁木齐 830012

中国人民大学应用统计科学研究中心,北京 100872

中国人民大学统计学院,北京 100872

展开 >

Expectile曲线 M分位数 伪分位数 函数型数据聚类 空气质量

北京市自然科学基金

1242005

2024

数理统计与管理
中国现场统计研究会

数理统计与管理

CSTPCDCSSCICHSSCD北大核心
影响因子:1.114
ISSN:1002-1566
年,卷(期):2024.43(2)
  • 24