I am fairly new to this topic but here is my problem:

I have stumbled across a paper (Robinson and Smyth, 2008) stating that the sample sum is a sufficient statistic for NB-distributed random variables.
I have tried to verify this by using the Fisher–Neyman factorization theorem $f_\theta(x)=h(x) \, g_\theta(T(x))$.
This is how far I have come: $\frac{\prod\Gamma(x_i+r)}{\prod x_i! \Gamma(r)^n}(1-p)^{n*r}p^{\sum{x_i}}$
It would be easy if it weren't for the Gamma function in the numerator as then $h(x)=\frac{1}{\prod x_i!}$ and $g_\theta(T(x))=\Gamma(r)^{-n}(1-p)^{n*r}p^{\sum{x_i}}$.

If I am on the wrong path, could somebody please help me solve this?


Are you given that $r$ is known? If it is, then you could define $h(x)$ as you have but including the product in the numerator. This would not be a problem as $r$ is known.

  • $\begingroup$ $r$ is being estimated in an Maximum-Likelihood Estimation approach. This is what irritates me. The sum can only be the sufficient statistic if $r$ is known but at the same time $r$ is being estimated with MLE. What am I not seeing here? $\endgroup$
    – Mark
    Oct 31 '12 at 10:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.