Talk:Cross-correlation matrix

function corr(..) used in definition is not defined, and there is no (at least i cannot find any) link to its definition in wikipedia.

ie. it reads: >>> For random variables X(s) and X(t) at different points s and t of some space, the correlation function is  C(s,t) = corr(X(s),X(t))

<<< but no explanation/link/definition of what "corr" is.

this is very important! it makes this page useless. --- I second the above comment, the definition given makes no sense. This page is useless. —Preceding unsigned comment added by 18.95.7.3 (talk) 23:14, 29 November 2008 (UTC)
 * I have added wikilink to article which now has definition of this notation. Melcombe (talk) 12:14, 2 December 2008 (UTC)
 * For me this page is unclear because it is unclear to me if X1 (one of the elements of X) is a scalar or a vector.
 * (if it is a vector, then I suppose the number of elements in Xj and Yk should be identical?) 83.83.238.65 (talk) 11:48, 1 July 2022 (UTC)

What then is the point of having an article titled "correlation function" which just points to another article for the definition of the correlation function? — Preceding unsigned comment added by 132.3.33.68 (talk) 19:06, 14 August 2012 (UTC)

I think a proper article should be self consistent. The introduction of the definition of correlation, just for the case under consideration, is definitely needed here. It would increase the length just by a couple of lines but the page would be clear. Stefano.nicotri (talk) 23:12, 26 November 2014 (UTC)stefano

Merger proposal
I am against merging in any of the other speciality-specific articles on correlation functions. Each speciality has its own conventions of terminology and don't necessarily agree on what the term "correlation function" means. Lets avoid confusion by keeping these different fields of application separate.Melcombe (talk) 16:45, 25 June 2008 (UTC)


 * I agree and have removed the merger template. However a merger with covariance function is worth considering. Skbkekas (talk) 03:56, 3 June 2009 (UTC)

Idea introduced by G.I Taylor
According to the historical notes in Fluid Mechanics (Kundu,Cohen), G.I Taylor introduced the idea of correlation function during a study of turbulent diffusion. Can anyone verify this and perhaps add this in the article ? —Preceding unsigned comment added by 193.157.204.95 (talk) 12:40, 27 March 2011 (UTC)

Disputed material
The recent drastic edits by "Fvultier" appear to give a definition of covariance rather than correlation, and then only if the expected values of the random vectors are zero. Maybe I'll alter it myself later. Michael Hardy (talk) 19:34, 29 December 2018 (UTC)


 * The terms correlation and covariance are unfortunately used different for things depending on the field and on the author of the book/paper/... Please check the sidebar on the top right of the article and let me know if you find any inconsistency in the definitions. I changed all the articles linked in the sidebar such that the term correlation refers to the expectation of a product (e.g. $$E[X \cdot Y]$$) and that covariance refers to the ecpectation of the product of the zero-mean random quantities (e.g. $$E[(X - \mu_X) \cdot (Y - \mu_Y)]$$).

Not to be confused with pearson product moment correlation
This should be put here.

Many fields use the term "correlation" to refer to pearson correlation. — Preceding unsigned comment added by 128.30.34.149 (talk) 16:28, 17 July 2019 (UTC)