A ready-to-use CGF object for the Negative Binomial distribution with
number of successes \(r\) and success probability \(p\).
The parameter_vector is \(c(r, p)\), and the actual
CGF is $$K(t) = r\,[\log p - \log(1 - (1-p)\,e^t)], \quad t < -\log(1-p).$$
Examples
NegBinCGF$K1(0, c(10, 0.25)) # E[X] = r(1-p)/p = 10 * 0.75 / 0.25 = 30
#> [1] 30