On Sep 2, 12:28 pm, Mark Dickinson <dicki...@gmail.com> wrote:
> On Sep 2, 2:51 pm, Thomas Philips <tkp...@gmail.com> wrote:
>
> > def student_t(df): # df is the number of degrees of freedom
> > if df < 2 or int(df) != df:
> > raise ValueError, 'student_tvariate: df must be a integer > 1'
>
> By the way, why do you exclude the possibility df=1 here?
>
> --
> Mark
I exclude df=1 hereBecause the variance is then infinite (in fact, the
distribution is then Cauchy). That said, your point is well taken;
allowing df=1 makes the Cauchy distribution available to users of
random, in much the same way as the Gamma makes the Chi-squared
available. Here's the revised code:
def student_tvariate(df): # df is the number of degrees of
freedom
if df < 1 or int(df) != df:
raise ValueError, 'student_tvariate: df must be a positive
integer'
x = random.gauss(0, 1)
y = random.gammavariate(df/2.0, 2)
return x / (math.sqrt(y/df))
I'll follow your suggestion, add in documentation and submit it to
bugs.python.com. Thanks for your guidance.
Thomas Philips
--
http://mail.python.org/mailman/listinfo/python-list
