Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 5 down vote favorite
1

Is there a nice limiting distribution of $\max( X_1,X_2,...,X_n) $ as $n$ goes to $\infty$, assuming that they are iid normal distributions with variance $\sigma^2$.

This is almost certainly a well known problem with a clever proof and nice solution, but I've been digging around and haven't found anything.

share|improve this question
2  
Rick Durrett's probability text has this as a fun problem. In the third edition, it's on page 83. – cardinal Aug 16 '12 at 10:02

2 Answers

up vote 5 down vote

With $M_n:= \mathrm{max}(X_1,\,X_2,\,\dots,\,X_n)$ it can be shown that $(M_n-b_n)/a_n$ is approximately Gumbel for some known $a_n>0$ and $b_n$. See http://www.panix.com/~kts/Thesis/extreme/extreme2.html and the herein quoted "example 1.1.7" from the book by de Haan and Ferreira: Extreme Value theory, an Introduction.

share|improve this answer
1  
+1 Great answer and a good book recommendation. There are several other good books on extreme value theory including the classic by Gumbel and the books by Galambos and the one book Leadbetter, Lindgren and Rootzen on the extension to stationary stochastic processes. A new and very readable recent book is the one by Stuart Coles. It is worth mentioning that the cumulative cdf for the Gumbel distribution exp(-e$^-$$^x$). – Michael Chernick Aug 16 '12 at 10:37
up vote 1 down vote

Check the book Tail Risk of Hedge Funds: An Extreme Value Application, chapter 3, section 3.1. They mention that the limiting distribution of the maxima follows either Gumbel, Frechet or Weibull distribution, whatever the parent distribution F.

share|improve this answer

Your Answer

 
discard

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.