Misplaced Pages

Gumbel distribution: Difference between revisions

Article snapshot taken from Wikipedia with creative commons attribution-sharealike license. Give it a read and then ask your questions in the chat. We can research this topic together.
Browse history interactively← Previous editNext edit →Content deleted Content addedVisualWikitext
Revision as of 13:04, 9 March 2004 editCharles Matthews (talk | contribs)Autopatrolled, Administrators360,239 editsm some format← Previous edit Revision as of 15:15, 25 July 2004 edit undoMulad (talk | contribs)Extended confirmed users17,648 edits try to do proper math modeNext edit →
Line 9: Line 9:
In particular, the Gumbel distribution is a special case of the '''Fisher-Tippett distribution''', also known as the '''log-Weibull distribution''', whose ] is In particular, the Gumbel distribution is a special case of the '''Fisher-Tippett distribution''', also known as the '''log-Weibull distribution''', whose ] is


:<math>p = \exp(-\exp((\mu-x)/\beta).</math> :<math>p = e^{-e^\frac{\mu-x}{\beta}}.</math>


The Gumbel distribution is the case where &mu; = 0 and &beta; = 1. The Gumbel distribution is the case where &mu; = 0 and &beta; = 1.


A more practical way of using the distribution could be A more practical way of using the distribution could be
:p=exp(-exp(-0.367*(A-x)/(A-M)) ;-.367=ln(-ln(.5)) :<math>p=e^{-e^{-0.367*\frac{A-x}{A-M}}} ;</math>
:<math>-.367=\ln(-\ln(.5))</math>
where M is the ]. To fit values one could get the median where M is the ]. To fit values one could get the median
straight away and then vary A until it fits the list of values. straight away and then vary A until it fits the list of values.


Its variates(ie to get a list of random values) can be given as ; Its variates(ie to get a list of random values) can be given as ;
:x=A-B*ln(-ln(rnd)) :<math>x=A-B*\ln(-\ln(rnd))</math>


Its percentiles can be given by ; Its percentiles can be given by ;
:x=A-B*ln(-ln(p)) :<math>x=A-B*\ln(-\ln(p))</math>


ie Q1=A-B*ln(-ln(.25)) ie <math>Q1=A-B*\ln(-\ln(.25))</math>


The median is A-B*ln(-ln(.5)) The median is <math>A-B*\ln(-\ln(.5))</math>


:Q3=A-B*ln(-ln(.75)) :<math>Q3=A-B*\ln(-\ln(.75))</math>


The mean is A+g*B where g =] = .57721. The mean is <math>A+g*B</math> where <math>g</math> = ] = .57721.


The sd = B * Pi()* sqr(1/6) The <math>sd = B * \pi() * \sqrt{\frac{1}{6}}</math>


Its mode is A Its mode is <math>A</math>




See also: ==See also==
* ] * ]

Revision as of 15:15, 25 July 2004

In probability theory and statistics the Gumbel distribution is used to find the minimum (or the maximum) of a number of samples of various distributions. For example we would use it to find the maximum level of a river in a particular year if we had the list of maximum values for the past ten years. It is therefore useful in predicting the chance that an extreme earthquake, flood or other natural disaster will occur.


The distribution of the samples could be of the normal or exponential type. The Gumbel distribution, and similar distributions, are used in extreme value theory.

In particular, the Gumbel distribution is a special case of the Fisher-Tippett distribution, also known as the log-Weibull distribution, whose cumulative distribution function is

p = e − e μ − x β . {\displaystyle p=e^{-e^{\frac {\mu -x}{\beta }}}.}

The Gumbel distribution is the case where μ = 0 and β = 1.

A more practical way of using the distribution could be

p = e − e − 0.367 ∗ A − x A − M ; {\displaystyle p=e^{-e^{-0.367*{\frac {A-x}{A-M}}}};}
− .367 = ln ⁡ ( − ln ⁡ ( .5 ) ) {\displaystyle -.367=\ln(-\ln(.5))}

where M is the median. To fit values one could get the median straight away and then vary A until it fits the list of values.

Its variates(ie to get a list of random values) can be given as ;

x = A − B ∗ ln ⁡ ( − ln ⁡ ( r n d ) ) {\displaystyle x=A-B*\ln(-\ln(rnd))}

Its percentiles can be given by ;

x = A − B ∗ ln ⁡ ( − ln ⁡ ( p ) ) {\displaystyle x=A-B*\ln(-\ln(p))}

ie Q 1 = A − B ∗ ln ⁡ ( − ln ⁡ ( .25 ) ) {\displaystyle Q1=A-B*\ln(-\ln(.25))}

The median is A − B ∗ ln ⁡ ( − ln ⁡ ( .5 ) ) {\displaystyle A-B*\ln(-\ln(.5))}

Q 3 = A − B ∗ ln ⁡ ( − ln ⁡ ( .75 ) ) {\displaystyle Q3=A-B*\ln(-\ln(.75))}

The mean is A + g ∗ B {\displaystyle A+g*B} where g {\displaystyle g} = Euler's constant = .57721.

The s d = B ∗ π ( ) ∗ 1 6 {\displaystyle sd=B*\pi ()*{\sqrt {\frac {1}{6}}}}

Its mode is A {\displaystyle A}


See also