多维信号非线性性的检验

Detecting nonlinearity in multidimensional signal

在多变量时间序列中引入了广义冗余的概念，给出了广义冗余与广义关联积分之间的关系，通过广义冗余与线性冗余之间的差别获得了定性检验多维信号非线性性的方法。一个线性自回归模型和Ｌｏｒｅｎｚ系统的仿真计算证实了这种方法的有效性。

In the paper, we introduce the generalized redundancy of multivariate time series, get the relationship of generalized redundancy and generalized correlation integration. The distinction of generalized redundancy and linear redundancy can be used to test the nonlineartiy of multidimensional signals qualitatively. The method is tested by a linear autoregression model and the Lorenz system.