Computing the CDF of the Borel Distribution

143 Views Asked by At

The pmf of the Borel distribution is given by

$$f(x)=\frac{e^{-x\rho } (x \rho )^{x-1}}{x!},\ \{x \in 1,2,\dots\}$$

I'm trying to compute the cdf, which is

$$F(x) = \sum_{j=1}^x \frac{e^{-j \rho } (j \rho )^{j-1}}{j!}.$$

Does anyone know of a closed form for the cdf of the Borel distribution?

========== UPDATE ====================

The closest I could find was this:

Haight, F. (1961). A Distribution Analogous to the Borel-Tanner. Biometrika, 48(1/2), 167-173.

In there, they have...

$$B(y,r)=\frac{r}{y} \binom{2 y-r-1}{y-1}$$

$$f(y)=B(y,r)\frac{\alpha ^{y-r}}{(\alpha +1)^{2 y-r}},\ y\in\{r,\dots,\infty\}, r \geq 1, r\in Z,$$

which has does have a cdf in terms of Hypergeometric functions.

But this distribution is NOT THE SAME as the Borel. I.e., they have the same mean, but different moments. I am unclear what they mean by ``analagous".