Multidimensional Cross-Recurrence Quantification Analysis (MdCRQA) – A Method for Quantifying Correlation between Multivariate Time-Series. (4th March 2019)

• Record Type: Journal Article
• Authors: Wallot, Sebastian
• Is Part Of: Multivariate behavioral research. Volume 54:Number 2(2019)
• Page Start: 173
• Page End: 191
• Publication Date: 2019-03-04
• Subjects: Multidimensional Cross-Recurrence Quantification Analysis -- multivariate time-series -- MdCRQA -- DCRP -- R -- MatLab; Psychometrics -- Periodicals; Psychology, Experimental -- Periodicals; Psychology, Experimental; Psychometrics; Periodicals; 150.15195
• DOI: 10.1080/00273171.2018.1512846
• Languages: English
• ISSNs: 0027-3171

Abstract

Abstract: In this paper, Multidimensional Cross-Recurrence Quantification Analysis ( MdCRQA ) is introduced. It is an extension of Multidimensional Recurrence Quantification Analysis ( MdRQA ), which allows to quantify the (auto-)recurrence properties of a single multidimensional time-series. MdCRQA extends MdRQA to bi-variate cases to allow for the quantification of the co-evolution of two multidimensional time-series. Moreover, it is shown how a Diagonal Cross-Recurrence Profile ( DCRP ) can be computed from the MdCRQA output that allows to capture time-lagged coupling between two multidimensional time-series. The core concepts of these analyses are described, as well as practical aspects of their application. In the supplementary materials to this paper, implementations of MdCRQA and the DCRP as MatLab- and R-functions are provided.