1
$\begingroup$

$X_1, X_2,\ldots,X_n$ is a sample from random variable $X$.

$$f(x\mid\theta)=\frac 1 {\sqrt{2\pi}}\exp{(-\frac{1}{2}(x-\theta)^2)}$$

I have to find a 1-dimensional sufficient statistic for $\theta$ and I'm not sure how to go about doing this. Any help would be appreciated.

I've got that the log-likelihood function is

$$\ell(\theta)=\sum^n_{i=1}(-\frac{1}{2}(x-\theta)^2)$$

$\endgroup$
1
$\begingroup$

Use the factorization theorem.

Hint #$1$: $$ \sum(x_i - \theta)^2 =\sum x_i^2-2\theta n\bar{x}_n+n\theta^2, $$ Hint #$2$:

The MLE is always a function of the MSS. What is the MLE of $\theta$?

$\endgroup$
0
$\begingroup$

$$ \prod_{i=1}^n f(x_i\mid\theta)= \prod_{i=1}^n \frac 1 {\sqrt{2\pi}}\exp\left(-\frac 1 2(x_i-\theta)^2\right) = \text{constant} \times \exp\left( - \frac 1 2 \sum_{i=1}^n (x_i-\theta)^2 \right). $$ So we examine that sum: \begin{align} \sum_{i=1}^n (x_i-\theta)^2 & = \sum_{i=1}^n ((x_i-\overline x)+(\overline x - \theta))^2 \quad \text{where } \overline x = \frac 1 n \sum_{i=1}^n x_i \\[10pt] & = \sum_{i=1}^n \Big((x_i-\overline x)^2 + 2(\overline x -\theta)(x_i-\overline x) + (\overline x - \theta)^2 \Big) \\[10pt] & = \left( \sum_{i=1}^n ((x_i-\overline x)^2) \right) + \underbrace{2(\overline x - \theta) \sum_{i=1}^n (x_i - \overline x)}_{\text{The factor $2(\overline x \, - \, \theta)$ can be} \\ \text{pulled out because it does not} \\ \text{change as $i$ goes from $1$ to $n.$}} + n(\overline x - \theta)^2 \end{align} But $\sum\limits_{i=1}^n (x_i-\overline x)=0,$ so this sum becomes $$ n(\overline x - \theta)^2 + \sum_{i=1}^n (x_i-\overline x)^2. $$ Now note that

  • one of these terms does not depend on $\theta$, and
  • the other one depends on $x_1,\ldots,x_n$ only through $\overline x.$

When you take the exponential, then instead of terms (i.e. things that are added), you have factors (i.e. things that are multiplied). Therefore Fisher's factorization theorem entails that $\overline X = (X_1+\cdots+X_n)/n$ is sufficient.

$\endgroup$

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.