|
From: | Frantisek Kluknavsky |
Subject: | Re: [Bug-gsl] Error in sample code |
Date: | Tue, 26 Feb 2013 17:57:17 +0100 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/20130219 Thunderbird/17.0.3 |
On 02/26/2013 05:32 PM, Rhys Ulerich wrote:
Hi Gerard,Hello, I think there is a bug in "doc/examples/stat.out" (GSL 1.9). "The estimated variance is 4.2984" should be "The estimated variance is 5.3730"Thank you for the report. The latest version of the source does contain 5.3730 (http://bzr.savannah.gnu.org/lh/gsl/trunk/annotate/head:/doc/examples/stat.out). Brian Gough corrected it back in 2010 (http://bzr.savannah.gnu.org/lh/gsl/trunk/revision/4662). I think you are looking at an older version of GSL. Version 1.9 is quite old. The last public release is 1.15. Cheers, Rhys
Hi,are you sure that "buggy" and correct behavior are not confused here? I do not know, but look at http://en.wikipedia.org/wiki/Bias_of_an_estimator and search for "usual definition of sample variance". If I understand it correctly, the "buggy" formula was unbiased, while now it is biased again.
Cheers, Fero
[Prev in Thread] | Current Thread | [Next in Thread] |