Dance Steps: A Statistical Quantification of Data Pair Similarities
9 Pages Posted: 12 Aug 2022 Last revised: 27 Mar 2024
Date Written: October 19, 2013
Abstract
Herein we attempt to quantify Tom McClellan’s notion of Dance Steps between two series of data, often of differing time periods (a current, or dependent and a previous, or independent period data series), looking for similarities in the turning points to use as a reference for the future turning points and direction of the current period data series. The visual recognition of dance steps requires a fuzzy relationship between the time difference and degree differences of turning points that loosely correspond in time. We discuss a means of measuring this on the same scale as correlation where perfect matching equates to a metric of +1, no match to 0, and perfect inverse matching to-1.
Keywords: Serial correlation, dependency
JEL Classification: C40, C49
Suggested Citation: Suggested Citation