摘要
应用复杂系统理论研究了汇率混沌数据延时嵌入相空间重构方法及其伪相图的获取技术.通过嵌入空间矩阵的分解来得到混沌时序的奇异谱曲线,依此获取混沌时序的内在本质特征及判定其噪声的组成比例,将混沌数据的除噪技术用于汇率数据非线性本质特征的提取.计算结果表明:伪相图在极大程度上反映了原复杂系统的内在本质特征.这一研究对于汇率混沌数据的除噪、本质特征获取、非线性混沌模型的确立技术等都有着极其重要的理论和实际意义.
With the theory of complicated system, this paper studies the method for restructuring the delayed embedded phase space of chaotic data of exchange rate, and the technology of obtaining its pseudo -phase portrait. Through decomposing the embedded space matrix, we can get the strange spectrum curve of chaotic time series, and can get its internal essential characteristics and determine the noise proportion of chaotic time series. Subsequently the method of reducing noise is used to extract nonlinear essential characteristics of exchange rate data. The calculation result indicates that the pseudo-phase portrait to the utmost extent reflects the internal essential characteristics of original complicated system. Therefore the above studies have very important significance to the technology of noise reduction, extracting essential characteristics and establishing the nonlinear chaotic model of exchange rate data.
出处
《系统工程学报》
CSCD
北大核心
2005年第6期620-624,共5页
Journal of Systems Engineering
基金
国家自然科学基金资助项目(70271071)
天津市教委资助课题(20041702)
关键词
混沌汇率波动数据
奇异谱曲线
伪相图
除噪技术
本质特征提取
chaotic vibration data of exchange rote
strange spectrum curve
pseudo-phase portrait
technology of noise reduction
essential characteristics extraction