answersLogoWhite

0

Think you've got this backwards. The exponential probability distribution is a gamma probability distribution only when the first parameter, k is set to 1. Consistent with the link below, if random variable X is distributed gamma(k,theta), then for gamma(1, theta), the random variable is distributed exponentially.

The gamma function in the denominator is equal to 1 when k=1. The denominator will reduce to theta when k = 1. The first term will be X0 = 1.

using t to represent theta, we have

f(x,t) = 1/t*exp(-x/t)

or we can substitute L = 1/t, and write an equivalent function:

f(x;L) = L*exp(-L*x) for x > 0

See:

http://en.wikipedia.org/wiki/Gamma_distribution

[edit]

To the untrained eye the question might seem backwards after a quick Google search, yet qouting wikipedia lacks deeper insight in to the question: What the question is referring to is a class of functions that factor into the following form:

f(y;theta) = s(y)t(theta)exp[a(y)b(theta)] = exp[a(y)b(theta) + c(theta) + d(y)]

where a(y), d(y) are functions only reliant on y and where b(theta) and c(theta) are answers only reliant on theta, an unkown parameter.

if a(y) = y, the distribution is said to be in "canonical form" and b(theta) is often called the "natural parameter"

So taking the gamma density function, where alpha is a known shape parameter and the parameter of interest is beta, the scale parameter. The density function follows as:

f(y;beta) = {(beta^alpha)*[y^(alpha - 1)]*exp[-y*beta]}/gamma(alpha)

where gamma(alpha) is defined as (alpha - 1)!

Hence the gamma-density can be factored as follows:

f(y;beta) = {(beta^alpha)*[y^(alpha - 1)]*exp[-y*beta]}/gamma(alpha)

=exp[alpha*log(beta) + (alpha-1)*log(y) - y*beta - log[gamma(alpha)]

from the above expression, the canonical form follows if:

a(y) = y

b(theta) = -beta

c(theta) = alpha*log(beta)

d(y) = (alpha - 1)*log(y) - log[gamma(alpha)]

which is sufficient to prove that gamma distributions are part of the exponential family.

User Avatar

Wiki User

14y ago

Still curious? Ask our experts.

Chat with our AI personalities

CoachCoach
Success isn't just about winning—it's about vision, patience, and playing the long game.
Chat with Coach
EzraEzra
Faith is not about having all the answers, but learning to ask the right questions.
Chat with Ezra
FranFran
I've made my fair share of mistakes, and if I can help you avoid a few, I'd sure like to try.
Chat with Fran

Add your answer:

Earn +20 pts
Q: How gamma are members of the exponential family prove?
Write your answer...
Submit
Still have questions?
magnify glass
imp