An observation $x$ is taken from a negative binomial distribution $X \sim \text{Negative-Binomial}(k,\pi)$. The parameter, $\pi$, is allocated a beta prior $\pi \sim (\alpha,\beta)$.

My attempt:

\begin{align} P(\pi \mid x) &\propto p(\pi)L(\pi) \\ P(\pi \mid x) &\propto \pi^{\alpha-1}(1-\pi)^{\beta-1} \pi^k (1-\pi)^x \\ P(\pi \mid x) &\propto \pi^{\alpha+k-1}(1-\pi)^{\beta+x-1} \end{align}

So the posterior is ${\rm Beta}(\alpha+k,\beta+x)$

Am I correct?

  • $\begingroup$ Welcome to our site! Please add the [self-study] tag & read its wiki. $\endgroup$ – Silverfish Apr 27 '16 at 0:01
  • $\begingroup$ @user164945 Why do you think you are not correct? What step are you doubtful about? $\endgroup$ – Greenparker Apr 27 '16 at 0:59
  • $\begingroup$ @Greenparker It's just I am not certain about my answer. $\endgroup$ – user164945 Apr 27 '16 at 1:03

Yes, that is correct. However, be aware that you can write the negative binomial in a different manner, so it may look different depending on which version you use. Looks like you're using the distribution defined by Wikipedia (https://en.wikipedia.org/wiki/Negative_binomial_distribution).


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.