0
$\begingroup$

Let $(X_1,..., X_n)$ be a random sample from a Poisson distribution where $X_i$'s are iid .Prove that $T=(\frac{n-1}{n})^{\sum_{i=1}^{n}X_i}$ is the MVUE for $q(\theta) = P(x=0)=e^{-\theta}$ and find its variance .

Based on Lehmann-Scheffe's theorem I can prove that $\sum_{i=1}^{n}X_i$ is a complete sufficient statistic since the joint density belongs to the exponential class of densities but I can not prove that $T$ is an unbiased estimator for $q(\theta)$ , all what I know that $\sum_{i=1}^{n}X_i \sim $ Poisson $(n\theta)$ . Any help or hint please ?

$\endgroup$
  • 2
    $\begingroup$ No need to shout in the title of your question using huge :) $\endgroup$ – Daniel W. Farlow May 24 '17 at 0:35
  • $\begingroup$ Are you sure the question is correct? T is not unbiased as $E[T]=E[\sum_{i=1}^{n}X_i](n-1)/n=n\theta(n-1)/n=(n-1)\theta $ $\endgroup$ – A-B-izi May 24 '17 at 0:50
  • $\begingroup$ @A.Brizzi It is raised to the power $\sum X_i $ $\endgroup$ – Bahgat Nassour May 24 '17 at 0:53
1
$\begingroup$

Look at the expectation of $T$, following the definition. We get: $E[T]=\sum_{k=1}^{\infty} (\frac{n-1}{n})^{k} e ^{-n\theta} \frac{n\theta^{k}}{k!}= e ^{-n\theta}\sum_{k=1}^{\infty} \frac{((n-1)\theta)^{k}}{k!} $

By definition of exponential,it follows that:

$E[T]=e ^{-n\theta}e ^{(n-1)\theta}=e ^{-\theta}$.

This is an unbiased estimator for $q(\theta)$.

Hope this helps

$\endgroup$
  • 1
    $\begingroup$ Check power of first term inside the first sum. $\endgroup$ – NCh May 24 '17 at 1:23
1
$\begingroup$

$$ \mathbb Ec^{\sum_{i=1}^n X_i} = \sum_{k=0}^\infty c^k \frac{(n\theta)^k}{k!}e^{-n\theta}=e^{cn\theta}e^{-n\theta}=e^{n\theta(c-1)}. $$ Substitute $c=\frac{n-1}{n}$ and conclude.

$\endgroup$

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.