0
$\begingroup$

Let $X_1,..,X_n$ be For a random sample $X_1,…,X_n$ from a Uniform$[\theta,θ+1] $distribution, with probability density function $f(x;\theta) = \begin{cases} 1, & \theta \le x \le \theta+1 \\ 0, & \text{otherwise} \end{cases}$ and $X_{(n)}=max(X_1,X_2,…,X_n)$ and $X_{(1)}=min(X_1,...,X_n)$. I want to show:

1)$X_{(1)}$ has the same distribution as $\theta - X_{(n)}$

2)$\hat{\theta}=X_{(1)}+X_{(n)}$ is an unbiased estimator of θ.

My work for 1) I know that $X_{(1)}$ has cdf $F_{X_{(1)}}(x)=1-(1-F(x))^n=\begin{cases} 0, & x\le\theta \\ 1-(1-(x-\theta))^n, & \theta< x<\theta+1 \\ 1, &\theta+1 \le x \end{cases}$.

$\endgroup$
  • $\begingroup$ $X_{(1)}$ does not have the same distribution as $\theta-X_{(n)}$, though I think $X_{(1)}- \theta$ has the same distribution as $\theta+1-X_{(n)}$. $\endgroup$ – Henry May 4 at 23:35
  • $\begingroup$ Similarly $X_{(1)}+X_{(n)}$ is not an unbiased estimator of $\theta$, though I think $\frac12(X_{(1)}+X_{(n)}-1)$ is an unbiased estimator of $\theta$. $\endgroup$ – Henry May 4 at 23:36
  • $\begingroup$ Just consider for example $\theta=7$. Then $7 \le X_{(1)} \le X_{(n)}\le 8$ while $-1 \le \theta-X_{(n)} \le 0$ and $14 \le X_{(1)}+X_{(n)} \le 16$ $\endgroup$ – Henry May 4 at 23:38
  • $\begingroup$ What you say is true for $U(0,\theta)$ but not $U(\theta,\theta+1)$. $\endgroup$ – StubbornAtom May 7 at 21:38
0
$\begingroup$

The numbers seem to be a little off. You can say, however, that $X_{(1)}$ is distributed like $2\theta + 1 - X_{(n)}$, while $\frac{1}{2}(X_{(1)} + X_{(n)} - 1)$ is an unbiased estimator of $\theta$ (as @Henry pointed out in the comments).

The first problem can be solved using a direct computation, but there's no need to expand any expressions or do any complicated algebra.

$$F_{X_{(1)}}(x) = 1 - (1 - F(x))^n.$$

as you noted. We can also compute,

$$F_{X_{(n)}}(x) = P(X_{(n)} < x) = P(X_k < x\text{ for } k=1,\dots,n) = (F(x))^n$$

Then,

$$F_{2\theta+1-X_{(n)}}(x) = P(2\theta + 1 - X_{(n)} < x) = P(X_{(n)} > 2\theta + 1 -x) = 1 - (F(2\theta+1-x))^n$$

However, we know that $F(x) = x - \theta$ for $\theta \leq x \leq \theta+1$, so

$$F(2\theta + 1 - x) = \theta + 1- x = 1 - (x - \theta) = 1 - F(x)$$

Conclude,

$$F_{2\theta+1-X_{(n)}}(x) = 1 - (F(2\theta+1-x))^n = 1 - (1 - F(x))^n = F_{X_{(1)}}(x).$$

For the second problem, just apply the first:

$$E[X_{(1)} + X_{(n)}] = E[2\theta + 1 - X_{(n)}] + E[X_{(n)}] = 2\theta + 1.$$

$\endgroup$

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.