Question

I have been using the scipy.stats.invgamma.rvs function to randomly select values from an inverse gamma distribution.

I’m struggling to find any information about the method Python/Scipy uses to generate these values (e.g. using the CDF or rejection sampling, transforming from another distribution etc.).

Does anyone know how this is undertaken? Your help would be very much appreciated! Apologies if I am repeating a question but I haven’t been able to find this information anywhere!

Answer 1

Scipy source code says: “Use basic inverse cdf algorithm for RV generation as default.”:

def _rvs(self, *args):
    ## Use basic inverse cdf algorithm for RV generation as default.
    U = mtrand.sample(self._size)
    Y = self._ppf(U,*args)
    return Y