[SciPy-dev] fix for ncx2 bug?

[Neal Becker]

Nathaniel Smith wrote:

> On Wed, Jun 17, 2009 at 12:00 PM, Neal Becker<ndbecker2 at gmail.com> wrote:
>> How about the boost reference?  This is an existence proof of code that
>> is supposed to be good quality.  Refer to non_central_chi_squared_cdf in
>> boost/math/distributions/non_central_chi_squared.hpp.
> 
> R's implementations of statistical functions may also a good
> reference; certainly they've been pounded on by more real
> statisticians than anything in scipy or boost. Also nice is that they
> always include citations in their help files -- just type
> help("pchisq") in R.
> 

R may be good too - I'm not familiar with it.

The boost ref has extensive test results including comparison against R.