In a uniform distribution where $0\leq X \leq \theta$, the pdf is represented as $f(X|\theta) = \frac{1}{\theta}I(0\leq X \leq \theta)$, and the likelihood is $L(\theta) = \prod\frac{1}{\theta}I(0\leq X \leq \theta) = \frac{1}{\theta^n}I(\max \ x_i < \theta, \min \ x_i > 0)$. This makes logical sense to me, as the $x_i$ must be smaller than $\theta$ and larger than 0.

What I don't understand is that I'm supposed to be able to logically conclude from this information that the MLE of $\theta$ is therefore the MAXIMUM of $x_i$. I am reading that the correct answer to the question of "what is the MLE of $\theta$?" for this question is that it is the MAXIMUM of $X$.

That doesn't make sense. Say that $x_1 = 1, x_2 = 2, x_3 = 3$, and these values are plugged in as $\theta$ in the likelihood function to see which of them maximizes our likelihood function. Then clearly 1 is the maximizer here, since for X=1, $\frac{1}{1^n}*I(true)=1$, whereas for X=3, $\frac{1}{3^n}*I(true)=0.333$ at MOST and is even smaller when n > 1, and that is all smaller than what we got for x = 1, so since x = 3 has a smaller likelihood, it wouldn't be the maximum likelihood estimator.

Explain where my logic here is wrong?


1 Answer 1


In your example $n=3$, $\min x_i = 1$, & $\max x_i = 3$. When $\theta=1$, $$I(\max x_i \leq \theta, \min x_i \geq 0)=I(3\leq 1,1\geq 0)=I(\mathit{false})=0$$ This factor, & therefore the likelihood, will be zero for any $\theta<\max x_i$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.