Talk:Canonical correlation

Mathematical notation
This and many other articles use mathematical notation but neither links to the article on mathematical notation, nor does that article explain the notation used on Wikipedia. I suggest our mathematical friends consider this challenge. --Lindorm 02:06, 7 August 2006 (UTC)


 * There is no special mathematical notation used on Wikipedia that is not also used elsewhere; it's just standard notation. I think this comment is silly.  It's like saying "This and many other articles use the English language but they do not link to the article titled English language."  Or "This and many other articles use graphics but there is no link to the article titled graphics." Michael Hardy 17:48, 7 August 2006 (UTC)


 * Frankly anyone without sufficient knowledge of the background subject matter to have already learnt the basic notation would never be able to follow an article on this topic anyway. It would be an unhelpful distraction to introduce irrelevant references to basic notation. JamesBWatson (talk) 08:38, 16 September 2009 (UTC)


 * I have to disagree. I have enough mathematical experience to perform what ever operation is needed, but it's been long enough since doing linear algebra work to have recognized some of the notation.  In particular transverse (') operations.  I'm going to add a minor subsection to reference and briefly explain some of the notation because not everyone who reads a page is an expert in the field that page falls under.141.219.198.171 (talk) 17:30, 13 February 2016 (UTC)

Column vectors
In the definition section, one should start with two data matricies rather than simple column vectors. You'd need a mXi matrix X and a corresponding nXi matrix Y to create a mXn cov(x,y) matrix. Drf5n (talk) 18:36, 17 March 2010 (UTC)
 * Well they are column vectors of the random variables. The actual sampled data would be given as two data matrices, and that's what you'd use to estimate the covariance matrix, but the notation is correct IIUC. I wonder if it would be clearer to make that step explicit in the definition, i.e. to describe the two sample matrices? --mcld (talk) 15:07, 1 April 2010 (UTC)
 * OK, added a sentence about this. Feel free to improve --mcld (talk) 15:18, 1 April 2010 (UTC)

Abbreviation
I didn't want to change this on the main page in case I'm mistaken, but, as I recall, the abbreviation for Canonical Correlation Analysis in CANCOR, whereas that for Canonical Correspondence Analysis is CCA. Napzilla (talk) 22:39, 25 October 2013 (UTC)

The term canonical should be explained
The word "canonical" is used in terms of algebraic theory of invariants -- as Hotelling (1936) clarified. — Preceding unsigned comment added by 31.183.229.101 (talk) 23:10, 2 January 2020 (UTC)

Example
The example is too trivial (n=m=1). — Preceding unsigned comment added by Merilius (talk • contribs) 11:29, 17 March 2021 (UTC)