Scribd
Upload
121 views3 pages

Inverse Gamma PDF

This document defines and summarizes properties of the inverse gamma distribution. It explains that the inverse gamma distribution is the reciprocal of a gamma distribution. The inverse gamma distribution has a density function defined for x > 0 that depends on shape parameter α and scale parameter β. Moments of the inverse gamma distribution are also derived, showing the mean is β/(α-1) for α > 1. Additionally, the inverse gamma distribution is the conjugate prior for the exponential distribution, so the posterior remains an inverse gamma distribution.

Uploaded by

unmsm1551
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
Download as pdf or txt
121 views3 pages

Inverse Gamma PDF

This document defines and summarizes properties of the inverse gamma distribution. It explains that the inverse gamma distribution is the reciprocal of a gamma distribution. The inverse gamma distribution has a density function defined for x > 0 that depends on shape parameter α and scale parameter β. Moments of the inverse gamma distribution are also derived, showing the mean is β/(α-1) for α > 1. Additionally, the inverse gamma distribution is the conjugate prior for the exponential distribution, so the posterior remains an inverse gamma distribution.

Uploaded by

unmsm1551
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
Download as pdf or txt
Download as pdf or txt
You are on page 1/ 3

Inverse Gamma Distribution

John D. Cook

October 3, 2008

Abstract

These notes write up some basic facts regarding the inverse gamma
distribution, also called the inverted gamma distribution. In a sense
this distribution is unnecessary: it has the same distribution as the
reciprocal of a gamma distribution. However, a catalog of results for
the inverse gamma distribution prevents having to repeatedly apply
the transformation theorem in applications.
Here we derive the distribution of the inverse gamma, calculate
its moments, and show that it is a conjugate prior for an exponential
likelihood function.

1 Parameterizations

There are at least a couple common parameterizations of the gamma distri-


bution. For our purposes, a gamma(α, β) distribution has density
1
f (x) = xα−1 exp(−x/β)
Γ(α)β α

for x > 0. With this parameterization, a gamma(α, β) distribution has


mean αβ and variance αβ 2 .
Define the inverse gamma (IG) distribution to have the density

β α −α−1
f (x) = x exp(−β/x)
Γ(α)

for x > 0.

1
2 Relation to the gamma distribution

With the above parameterizations, if X has a gamma(α, β) distribution


then Y = 1/X has an IG(α, 1/β) distribution. To see this, apply the
transformation theorem.


d
y −1

fY (y) = fX (1/y)
dy
1
= y −α+1 exp(−1/βy) y −2
Γ(α)β α
(1/β)α −α−1
= y exp(−(1/β)/y)
Γ(α)

3 Moments

Next we calculate the moments of X ∼ IG(α, β). If α > n,


βα ∞
Z
E(X n ) = xn x−α−1 exp(−β/x) dx
Γ(α) 0
βα
Z ∞
= xn−α−1 exp(−β/x) dx
Γ(α) 0
β α Γ(α − n)
=
Γ(α) β α−n
β n Γ(α − n)
=
(α − 1) · · · (α − n)Γ(α − n)
βn
= .
(α − 1) · · · (α − n)

In particular, for α > 1


β
E(X) =
α−1
and for α > 2
β2
E(X 2 ) =
(α − 1)(α − 2)
and so for α > 2
β2
V ar(X) = E(X 2 ) − E(X)2 = .
(α − 1)2 (α − 2)

2
4 Conjugate prior for exponential likelihood

Finally, suppose than an observation X | µ ∼ exponential(µ) and a priori


µ ∼ IG(α, β). By Bayes’ theorem, the posterior distribution on µ given an
observation X = x is proportional to
1 1 1
exp(−x/µ) α+1 exp(−β/µ) = α+2 exp(−(β + x)/µ).
µ µ µ

When normalized to be a probability distribution, the result is an IG(α + 1,


β + x) distribution. In general, after observing x1 , x2 , . . . , xn , the posterior
distribution on µ is IG(α + n, β + ni=1 xi ).
P

The motivation for parameterizing the inverse gamma distribution the


way we do is to make the posterior distribution have the simple form above.

This document available at http://www.johndcook.com/inverse gamma.pdf.

You might also like