This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: gsl-chisquare cumulative density function
- To: "Varghese John" <vjohn111 at hotmail dot com>
- Subject: Re: gsl-chisquare cumulative density function
- From: Brian Gough <bjg at network-theory dot co dot uk>
- Date: Thu, 27 Sep 2001 16:12:49 +0100 (BST)
- Cc: gsl-discuss at sources dot redhat dot com
- References: <F188NuhsZnOXOweys1S0000b058@hotmail.com>
Varghese John writes:
>
> Hi,
> Not sure if this will be of help, but here are three solutions.
>
> 1)the Normal CDF is nothing but
> the error function upto normalisation.
> So you can use the gsl Error function : gsl_sf_erfc.
>
> 2) Use a simple rational approx.
>
> 3)Use the numerical Integration.
Right, ideally I would like to see a separate cdf/ directory with
functions for direct computation of both P(x')=integ(p(x),-inf,x') and
Q(x)=1-P(x) to full precision. A lot of the cases are already
available in specfunc.