Gamma Distribution Confidence Interval

[jaycool1995]

How would you go about finding the confidence interval for the parameters of the gamma distribution? I have had a look online and haven't found anything with the answer...
Thanks

 

[SW VandeCarr]

You can get good approximations to the CI of the gamma distribution (which is a two parameter exponential distribution) from the normal, Poisson, or (inverse)chi square approximations. There are also exact methods. This paper discusses all of these methods:

http://wonder.cdc.gov/wonder/help/cancer/FayFeuerConfidenceIntervals.pdf

 

[jaycool1995]

So can i use another distribution to approximate the parameters (e.g the mean) of the gamma dist?
Thanks

 

[SW VandeCarr]

So can i use another distribution to approximate the parameters (e.g the mean) of the gamma dist?
Thanks

Yes. The gamma distribution morphs from Poisson like to normal like "shapes" depending on the parameter k. For k more than 3, the normal approximation is good. "k" relates to the failure or waiting times. k's value can be taken from the PDF where the power of the variable x is k-1.