Talk:Noncentral chi-squared distribution
|WikiProject Statistics||(Rated B-class, Low-importance)|
|WikiProject Mathematics||(Rated B-class, Low-importance)|
The function presented as the distribution function is in fact the probability density function.
220.127.116.11 21:43, 19 October 2007 (UTC) Dennis Clason New Mexico State University Statistics Center
I've added: a motivation section in the Introduction (to address the motivation suggestion by User:Geometry_guy given some three years ago); normal approximation to the non-central chi-square; implementation in Stata software; additional reference to Muirhead (2005), one of the highest rigor multivariate statistics textbooks. Enjoy :). I would boost the quality to B+ class; I agree it's low priority.
I've moved the 'Related Links' section to 'Software', and added examples, commands and links for a few packages. This seems preferable to arguments over who's online calculator should be used, which have plagued some distribution pages.
Anyone who feels slighted because their favourite software isn't included should simply add another line to the table.
A second point is to provide a few benchmark examples, to avoid confusion over the different definitions of the noncentrality parameter in different references.
I added an online calculator, which is also a wiki page welcoming collaboration. Online calculators are desperately needed by students and colleagues who have never written any line of code. I'd like Wikipedia can develop or adopt some extension like R Extension for MediaWiki, and support interactive online calculating and plotting.
A compromised approach could be linking R scripts to a contributed Rweb server, which provides Web based interface to R. Example: qchisq(0.5,3,1.5)
- Concerning the information you added to the article, there are quite a few websites which do similar calculations. I think it's better to stick to the most commonly used programs (Matlab and R), as most readers of this page probably do not know what MediaWiki is. --Zvika (talk) 06:15, 3 November 2008 (UTC)
- Fair enough. I've removed the reference to the R extension for MediaWiki since it really has nothing to do with noncentral chi-square. I've left the link to the online calculator as I agree it can be useful. What do you think? --Zvika (talk) 18:08, 6 November 2008 (UTC)
Zero degrees of freedom
I have added a ref/citation for the case of zero degrees of freedom. I think most of the formulas already here are valid for this case (except possibly the density, depending on interpretation) and not just for k>0, but I cannot check this. Melcombe (talk) 17:14, 30 October 2009 (UTC)
Closed Form Approximation
The formula for the closed for approximation for the CDF is syntactically incorrect (mismatched parentheses). I would have fixed it, but the referenced article can be viewed only with a $25 fee. Could someone who knows what the approximation really is please fix it. And, on that subject, it would be nice to cite a freely available source in addition to the primary reference. DMJ001 (talk) 03:36, 5 May 2010 (UTC)
- I have corrected the formula to match what needed to be an earlier source, and I have also added another book source that has the formula, which may be easier to obtain. Melcombe (talk) 12:27, 6 July 2010 (UTC)
Characteristic function reference
The reference  proving the formula for the characteristic function contains a flawed parity argument. I would suggest removing or changing the reference until its author fixes his proof. --18.104.22.168 (talk) 14:24, 18 November 2011 (UTC)
I feel that the formula for noncentral chi squared only apply in the case of normal distribution with unit variance. Think of a squared sum of distribution N(0, sigma_i), then, using the current formula, lambda would be equal to zero, and we will be back to a central chi-squared. This is incoherent. — Preceding unsigned comment added by 22.214.171.124 (talk) 12:16, 1 October 2014 (UTC)