摘要
收敛交叉映射(CCM)是一种分析非线性系统中时间序列变量间因果关系的方法。其不同于传统的线性系统分析方法,是通过对变量进行状态空间重构来获取变量的历史信息,随着时间序列不断增长,当其估计性能呈现收敛的性质时,可以判断因果关系的存在。本文介绍了CCM的发展史及其较传统的格兰杰因果检验的优点,详细阐明了CCM的原理、算法过程和实现途径。CCM作为一种针对变量间具有弱到中等强度耦合关系的系统分析方法,可以用来有效地解决非线性生态系统多变量间复杂的因果关系问题。将该方法应用于具有空间信息的多点位时间序列变量间因果分析时,应充分考虑点位间的空间自相关性,与可以去除变量及序列间空间相关性的方法相结合,从而确保CCM对变量因果关系的分析更加准确,结果也更具有信服力。
The convergent cross mapping(CCM)is a method to analyze causality of nonlinear time series variables.Different from the traditional linear system analysis method,CCM gets historical information based on their state space reconstruction.The presence of causality can be confirmed when the estimated values perform convergent with time series extension.Here,we introduced the development history of CCM and its advantages over the traditional Granger causality test,and elaborated the principle,algorithm process,and implementation approach.As a system analysis method aiming at the coupling relationship between variables from weak to moderate,CCM can effectively solve the complex causality among nonlinear multivariable in ecosystems.When it is applied to the causality analysis of multi-point time series variables with spatial information,the spatial autocorrelation among points should be fully considered and combined with the method that can remove the spatial correlation between variables and sequences,so as to ensure more accurate causality analysis using CCM and more convincing results.
作者
王丹雨
朱媛君
杨晓晖
WANG Dan-yu;ZHU Yuanjun;YANG Xiao-hui(Institute of Desertification Studies,Chinese Academy of Forestry,Beijing 100091,China)
出处
《应用生态学报》
CAS
CSCD
北大核心
2021年第12期4539-4548,共10页
Chinese Journal of Applied Ecology
基金
国家自然科学基金国际(地区)合作与交流项目(32061123005)和国家自然科学基金项目(41971061)资助。
