Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $X_{1},X_{2},...,X_{n}$ represent a random sample from a distribution with pdf:

$f(x; \theta)=e^{-(x-\theta)}, \theta \le x<\infty, -\infty<\theta<\infty$ | zero elsewhere

I need to find the MLE $\hat {\theta}$ of $\theta$.

Since the support space of the pdf is dependent on $\theta$, do I need to express the pdf in terms of an indicator function?

i.e. $f(x; \theta)=e^{-(x-\theta)}I_{(\theta,\infty)}(x)$

If so, do I find the MLE in the standard manner?

i.e. $L(x;\theta)=\displaystyle \prod^{n}_{i=1} f(X_{i};\theta)=e^{-(\sum^{n}_{i=1}X_{i}-n\theta)}I_{(\theta,\infty)}(X_{(1)})$

$\ln L(x;\theta)=-\displaystyle \sum^{n}_{i=1} X_{i} +n\theta +\ln I_{(\theta,\infty)}(X_{(1)})$

The next step would be to take the partial derivative of the log-likelihood function with respect to $\theta$, but how would I find the partial derivative of the indicator function?

Am I approaching this question in the correct manner? Any help would be greatly appreciated!

share|cite|improve this question
You can't write the log-likelihood like this. Take a look at your likelihood function instead. Which value of $\theta$ would maximize this function? – Patrick Li Jan 29 '13 at 5:51
Partial derivatives are not the only way to maximize a function. Try to rewrite $L$ as $L(x;\theta)=c\mathrm e^{n\theta}\mathbf 1_{\theta\lt y}$ and to think about the shape of the function $\theta\mapsto c\mathrm e^{n\theta}\mathbf 1_{\theta\lt y}$. – Did Jan 29 '13 at 8:43
up vote 1 down vote accepted

Note that the likelihood function is $$\begin{align}L(\theta |x) &=e^{-(\sum^{n}_{i=1}x_{i}-n\theta)},x_i \geq \theta,\forall i \\ &=e^{-\sum^{n}_{i=1}x_{i}}\cdot e^{n\theta},x_{(1)} \geq \theta \end{align}$$ Now note that $L(\theta |x)$ is maximum

iff $e^{-\sum^{n}_{i=1}x_{i}}\cdot e^{n\theta}$ is maximum subject to the restriction $x_{(1)} \geq \theta $

iff $e^{n\theta}$ is maximum subject to the restriction $x_{(1)} \geq \theta $

iff $\theta =x_{(1)}$.

Here Your MLE is $x_{(1)}=min(x_1,\dots ,x_n)$.

share|cite|improve this answer
how do we reach from the second last step and conclude that $\theta$ = x1 – statBeginner Oct 31 '14 at 4:00

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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