0
$\begingroup$

Let $X_1, \ldots, X_n$ be a random sample of size $n$ from a distribution having the pdf $$f(x)=\frac{2x}{\theta^n}.$$ I need to find the MLE for the mean of the distribution but am not sure how.

$\endgroup$
0
$\begingroup$

The idea behind MLE is to pick the value of the parameter to maximize the chances that the currently observed sample would be generated. Assuming $\{X_k\}$ are iid and actually result in the values $\{x_k\}$, we have for $\epsilon \to 0$ $$ \begin{split} p(\theta) &= \mathbb{P}\left[ X_1 \in [x_1 - \epsilon, x_1 + \epsilon], \ldots, X_n \in [x_n - \epsilon, x_n + \epsilon] \right] \\ &= \prod_{k=1}^n \mathbb{P}\left[ X_k \in [x_k - \epsilon, x_k + \epsilon] \right] \\ &= \prod_{k=1}^n f(x_k) = \prod_{k=1}^n \frac{2 x_k}{\theta^n} = \left(\frac{2}{\theta^n}\right)^n \prod_{k=1}^n x_k. \end{split} $$ It helps to maximize $$ L(x) = \ln f(x) = n \ln \left(\frac{2}{\theta^n}\right) + \sum_{k=1}^n \ln(x_k) $$ with respect to $\theta$ getting the best value for $\theta$ in terms of the $x_k$'s...

$\endgroup$

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.