基于谱系聚类的全球各国新冠疫情时间序列特征分析
谢聪慧(1997— ),女,硕士生,主要从事遥感与地理信息系统应用。E-mail: xieconghui19@mails.ucas.ac.cn |
收稿日期: 2020-08-17
修回日期: 2020-11-21
网络出版日期: 2021-04-25
基金资助
中国科学院战略性先导科技专项(A类)(XDA23100000)
国家科技基础资源调查专项(2017FY101004)
国家自然科学基金项目(42041001)
版权
Analysis of Time Series Features of COVID-19 in Various Countries based on Pedigree Clustering
Received date: 2020-08-17
Revised date: 2020-11-21
Online published: 2021-04-25
Supported by
Chinese Academy of Sciences strategic leading science and technology project (Class A)(XDA23100000)
National Science and technology basic resources survey project(2017FY101004)
National Natural Science Foundation of China(42041001)
Copyright
COVID-19暴发以来,世界各国疫情呈现出不同的时序特点,研究不同国家疫情发展模式的特点,揭示其背后的主导因素,可为未来防控策略提供参考。为了揭示不同国家疫情时间序列之间的异同,本文提取了主要疫情国家每日新增病例时间序列的标准差、Hurst指数、治愈率、增长时长、平均增长率、防控效率进行谱系聚类,并从经济、医疗、人文冲突方面对聚类结果进行了成因分析。结果表明,全球疫情发展模式可分为3大类:C型、S型和I型。C型国家时间序列的特点是持续波动上涨,治愈率较低,原因是其人文冲突不利于疫情防控,经济医疗资源经过长时间大量消耗已趋于匮乏,建议在防控中加强宣传疏导,改变观念,统筹分配经济、医疗资源;S型国家时间序列的特点是快速上升后立即下降,并最终保持稳定趋势,总体治愈率较高,其原因是这类国家国内稳定,经济医疗水平较高,以及防控措施及时,建议加强国际合作和科学研究,并为可能到来的二次疫情做好准备;I型国家时间序列特点是缓慢上涨,整体发展趋势不稳定,治愈率较低,原因是其暴发比较晚,程度较小,大部分经济医疗水平以及人文冲突不利于疫情防控,建议汲取较好的防控经验,实施严格的隔离措施,尽量满足疫情期间物资需求,优化治疗方法。
谢聪慧 , 吴世新 , 张晨 , 孙文涛 , 何海芳 , 裴韬 , 罗格平 . 基于谱系聚类的全球各国新冠疫情时间序列特征分析[J]. 地球信息科学学报, 2021 , 23(2) : 236 -245 . DOI: 10.12082/dqxxkx.2021.200470
Since the outbreak of COVID-19, countries around the world have shown different time-series characteristics. Studying the characteristics of the development patterns of different countries and revealing the dominant factors behind them can provide references for future prevention and control strategies. In order to reveal the similarities and differences between the epidemic time series in different countries, this article extracts the standard deviation, Hurst index, cure rate, growth time, average growth rate, and prevention and control efficiency of the daily time series of new cases in the main epidemic countries for pedigree clustering. We also analyzes the causes of clustering results from the aspects of economics, medical treatment, and humanistic conflicts. The results show that the global epidemic development model can be divided into three categories: C-type, S-type, and I-type. The time series of C-type countries are characterized by continuous fluctuations and rising, and the cure rate is low. The reason is that humanistic conflicts are not conducive to epidemic prevention and control. Economic and medical resources have become scarce after a long period of large consumption. It is recommended to strengthen publicity and guidance in prevention and control, change concepts, and coordinate the allocation of economic and medical resources. The time series of S-type countries is characterized by a rapid rise and then an immediate decline, and eventually maintains a stable trend. The overall cure rate is relatively high. The reason is that these countries have domestic stability, high economic and medical standards, and timely prevention and control measures. It is recommended to strengthen international cooperation and scientific research, and prepare for the possible second epidemic. The time series of I-shaped countries is characterized by a slow rise, the overall development trend is unstable, and the cure rate is low. The reason is that its outbreak is relatively late and less severe. Most of the economic and medical levels and humanistic conflicts are not conducive to epidemic prevention and control. It is recommended to learn better prevention and control experience, implement strict isolation measures, try to meet the material needs during the epidemic, and optimize treatment methods.
表1 COVID-19综合防控严峻指数构成Tab.1 Composition of COVID-19 comprehensive prevention and control severity index |
评价指标 | 评价内容 | 意义 | 来源 |
---|---|---|---|
经济因子 | 经济实力、增长、发展 | 经济基础决定了国家是否有足够的经济实力支撑抗疫进度,但经济发达也意味着人群活动性强,从而加大疫情传播,影响时间序列发展趋势 | 兰德公司传染病脆弱性指数中的经济指标[26] |
交通运输、技术、通讯等基础设施 | |||
医疗因子 | 个人医疗服务的获取和质量 | 国家的医疗体系是否完善决定了治愈率和死亡率的大小 | 《柳叶刀》发布的2019全球医疗质量和可及性榜单[27] |
人文冲突因子 | 暴力内部冲突概率 | 国家的人文冲突程度影响着疫情的传播,国内冲突较多会增强疫情的传播,国民配合程度也会较低,从而降低防控效率,新增确诊病例时间序列也会难以下降 | 欧盟委员会(JRC)联合研究中心开发的多危害风险评估信息全球风险指数增强版(GRI)中的人文指标[28] |
高暴力内部冲突概率 | |||
国家权力冲突强度 | |||
国家以下各级冲突强度 |
图2 各国疫情时间序列特征谱系聚类结果注:纵坐标的数字代表国家序号。 Fig. 2 The pedigree clustering results of the time series characteristics of epidemics in various countries |
[1] |
|
[2] |
|
[3] |
|
[4] |
|
[5] |
|
[6] |
|
[7] |
|
[8] |
|
[9] |
|
[10] |
|
[11] |
|
[12] |
|
[13] |
|
[14] |
|
[15] |
|
[16] |
|
[17] |
|
[18] |
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
[23] |
|
[24] |
宋辞, 裴韬. 基于特征的时间序列聚类方法研究进展[J]. 地理科学进展, 2012,31(10):1307-1317.
[
|
[25] |
|
[26] |
Moore, Melinda,
|
[27] |
|
[28] |
|
[29] |
袁勇. 非洲疫情形势值得密切关注[N]. 经济日报, 2020-04-11(002).
[
|
[30] |
刁海洋. 全球疫情复杂演化关键问题依然待解[J]. 健康中国观察, 2020(4):90-91.
[
|
[31] |
|
[32] |
|
/
〈 |
|
〉 |