-1
$\begingroup$

Let $X_1,\ldots,X_m$ be i.i.d. having the uniform distribution $U(0, \theta_x)$ and $Y_1,\ldots, Y_n$ be i.i.d. having the uniform distribution $U(0, \theta_y)$. Suppose that $X_i$’s and $Y_j$’s are independent and that $\theta_x > 0$ and $\theta_y > 0$. Find the UMVUE of $\theta_x/\theta_y$ when $n > 1$.

I don't know how to proceed with this exercise and I'd like some help.

I do know that the UMVUE of a unfiorm distirbution is $(n+1)X_{(n)}/n$ but I think that this isn't the way to do this exercise.

Thanks.

$\endgroup$
  • 2
    $\begingroup$ The title doesn't correctly represent what the question asks. $\endgroup$ – Glen_b Dec 17 '18 at 6:21
1
$\begingroup$

First, one only need look at sufficient statistics:

The sufficient statistic for a Uniform sample over $(0,\theta)$ is $X_{(N)}$ Hence a sufficient statistic for this problem is $(X_{m:m},Y_{n:n})$

Second, one need find an unbiased estimator based on a sufficient statistic:

Since$$\mathbb{E}_{\theta_x}[X_{m:m}]=\frac{m}{m+1}\theta_x\ \text{ and }\ \mathbb{E}_{\theta_y}[Y_{n:n}^{-1}]=\frac{n}{n-1}\theta_y^{-1}$$ an unbiased estimator of $\theta_x\theta_y^{-1}$ is $$\frac{m}{m+1}\frac{n-1}{n}X_{m:m}Y_{n:n}^{-1}$$

Last, one can call for a completeness argument.

| cite | improve this answer | |
$\endgroup$

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.