度量两个序列非线性相关性的一种方法
MEASURING NONLINEAR DEPENDENCE BETWEEN TIME SERIES
-
摘要: 基于分数维数,提出了非线性相关度的概念,用于度量两列经济数据的非线性相关程度,以解决非线性经济预测中的变量选择问题.仿真结果与应用结果说明该方法效果较好.Abstract: In this paper, a statistic-Nonlinear Dependence Coefficient based on fractal dimension is proposed. This statistic is designed for measuring nonlinear dependence between time series, and can be used in variable selecting for economical forecasting. Numerical results show the statistic is reasonable.