Talk:Total sum of squares

Latest comment: 12 years ago by Melcombe in topic Sum of squares vs. variance

Suggestion that this article be merged with Least squares

edit

Keeping this article separate makes it easier to check the definition when clicking on a link from another article (e.g. Coefficient of variation) that references the definition.  —Daphne A 09:55, 7 August 2006 (UTC)Reply

I agree, this way its more helpful. Merging it will reduce the chance of someone finding exactly what they want --Now might i do it pat 20:50, 6 January 2007 (UTC)Reply

This needs to be merged since total sum of squares is a result from least squares regression analysis--Vtodorov3 01:42, 10 November 2006 (UTC)Reply


I looked it up while specifically looking for Total sum of squares. Bravo for it being so easy on wiki and not being merged with an OLS article. Especially since it applies to many other statistical concepts. What if you look it up in the context of having read an article about gls?


I agree with not merging them. The total sum of squares is used for a variety of different tasks beyond regression analysis. Removing merge proposal. -- DSGruss 05:21, 28 March 2007 (UTC)Reply

Sum of squares vs. variance

edit

This should say something about how partitioning of sum of squares into explained + residual relates to the idea that total variance is sum of variances of individual contributing factors. Sum of squares is not variance, because it is unscaled. So, the tradition of saying that a certain fraction of variance is explained by a specific treatment factor seems wrong. — Preceding unsigned comment added by Tedtoal (talkcontribs) 04:35, 22 July 2012 (UTC)Reply

There is already an article ... Partition of sums of squares ... which may be doing part of what you want, and it is linked from this article by the redirect from Sum of squares (statistics). Melcombe (talk) 08:12, 22 July 2012 (UTC)Reply