期刊
INTERNATIONAL JOURNAL OF SYSTEMS SCIENCE
卷 38, 期 1, 页码 47-60出版社
TAYLOR & FRANCIS LTD
DOI: 10.1080/00207720601014552
关键词
nonlinear correlation tests; CCF; MIMO models; higher order CCF; NARMAX models
A set of novel correlation tests using omni-directional cross-correlation functions (ODCCFs), which are based on the first order cross-correlation functions (CCF), are proposed in the present study to comprehensively detect nonlinear relationships between variables. Then the ODCCFs are combined into a set of concise formulations to provide better illustration of detected correlations and reduce the number of correlation plots. Compared to the other approaches, the new methodology brings much more power in detection of nonlinear correlations. The efficiency and effectiveness of the new algorithm are demonstrated through simulation studies and comparisons with other linear and nonlinear correlation tests. The results can be widely applied in many relevant fields.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据