复杂系统中统计相关性测量的熵方法研究  

Study of entropy method of measurement for statistical correlation in complex system

作　　者：赵亚丽[1] 西广成[1] 易建强[1] 

机构地区：[1]中国科学院自动化研究所复杂系统与智能科学重点实验室,北京100080

出　　处：《系统工程学报》2005年第4期427-431,共5页Journal of Systems Engineering

基　　金：国家973计划资助项目(2003CB517106);国家自然科学基金资助项目(90209034).

摘　　要：在熵方法定义的关联度系数法的基础上,提出了一种计算复杂系统统计相关的广义关联度系数法,并将其应用到生态经济区划和中医理论(齿痕舌出现的特点)中,进而证明了该方法的有效性.结果表明:广义关联度系数法在计算时不需要特定的数据,从而减小了数据的不确定性和计算量;既适用于线性的情况也适用于非线性的情况;复杂系统分类所需的判据减少,使复杂系统分类简单化;可以对系统变量(或子系统)之间的相关性进行定量的描述.A method of generalized-relating-degree coefficient is presented to measure statistical correlation in complex system, wich is based on the method of original-relating-degree coefficient defined by entropy method. The proposed method is applied to the regionalization of ecological economy and the theory analysis of traditional Chinese medicine with characteristics of the appearance of teeth-printed tongue. The method does not need specific sample data, which reduces data uncertainty and computation quantity. It can be applied to linear and nonlinear problems. Also, it simplifies the classification of complex systems due to the reduction of classification criterions. Moreover, it can describe the relation among variables or sub-systems of system quantitatively.

关 键 词：复杂系统 广义关联度系数法 熵方法 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置] R241.25[自动化与计算机技术—控制科学与工程]