From bug-octave-request at bevo dot che dot wisc dot edu Tue Jul  8 08:22:31 2003
Subject: Re: normal_inv behavior
From: =?iso-8859-1?Q?Mikko_Lepp=E4nen?= <mle at e240a dot w3 dot ton dot tut dot fi>
To: Claude Gamache <jfcgamache at sympatico dot ca>
Cc: bug-octave at bevo dot che dot wisc dot edu
Date: Tue, 8 Jul 2003 00:46:11 -0500

Hi,

On Mon, Jul 07, 2003 at 10:12:45PM -0500, Claude Gamache wrote:
> I am getting what I think to be a bug with the function normal_inv.

Fortunately this is not the case.

> When you run the following lines:
> 
> dist_sigma      = 0.7;
> dist            = normal_inv( dist_rand, dist_mean, dist_sigma);
> dist_calc_std   = std(dist)
> 
> I get the following results:
> dist_calc_std = 0.82814
> 
> But the standard deviation should be very close to 0.7, not 0.82...
> 
> I get the same kind of error with dist_sigma=3.5
> dist_calc_std = 1.8623   <---- this should be very close to 3.5...
> 
> Am I doing something wrong or is this really a bug?

It seems that you have slightly missread the documentation for
normal_inv. The third parameter of the function is the variance, not the
standard deviation. As you probably know, variance is the square of
standard deviation, so you should probably call
normal_inv( dist_rand, dist_mean, dist_sigma.^2);

And indeed, 0.82814^2 ~ 0.686, and 1.8623^2 ~ 3.47.

Hope this helps,
-- 
Mikko Leppänen



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------