stat: Add binomial distribution to distuv
Created by: deasmhumhna
I considered opening an issue first, but then I just went ahead and wrote the code. This code follows the outline in "NUMERICAL RECIPES IN C: THE ART OF SCIENTIFIC COMPUTING". The Poisson approximation for np < 1
was extended to rejection sampling due to failed tests for even moments (variance and ex-kurtosis). 2.6e-02
is a bit of a magic number; it's close to the (smooth) maximum of the difference in log-likelihood between the Poisson and binomial for n > 25
.