An alternative measure of positive correlation for bivariate time series

Abstract

Nadarajah et al. (2018) have recently proposed an alternative measure for positive correlation between two variables of interest say X and Y. This takes into account the idea of concordance or association between the two variables. They have made use of extensive illustrations including a series of confidence intervals to show the applicability and interpretation of their proposed measure. In this work, we propose an alternative measure for positive correlation between integrated time series processes. We support this idea through simulated examples and real data illustrations from climatology and finance.

Keywords:

• Auto regressive
• bivariate
• correlation
• Non-stationarity
• Time Series

Subject Classifications:

• 37M10; 62M10; 91B84