[R] Anderson-Darling for Gamma Distribution, and Extreme Value Distribution

[ann young]

   Any  examples  of implementing Anderson-Darling for Gamma Distribution
   quickly:  that  is, getting critical values for different sample sizes
   for  different values of the Gamma distribution parameters in order to
   do  the  Goodneess  of  fit test?  These appear difficult to get since
   there   are   an   infinte  number  of  values  for  the  distribution
   parameters.   To  run this test on many models (e.g., say each traffic
   light  at  an  intersection in a city for 1 minute at the start or the
   common lunch hour of 11:30 AM each day), takes too long to simulate to
   get the critical values.  Looking for appoximations so that when I the
   mle  estimates  of  the  distribution parameters for Gamma, I can then
   do a  quick  Goodness  of  fit test to see if the data distribution is
   Gamma.   I have quick search tables or approximating equations for the
   other  common  distbutions  (Exp,  Norm,  Weibull,  ect.), but not the
   Gamma.
     _________________________________________________________________

   [1]Need a break? Find your escape route with Live Search Maps.

References

   1. http://g.msn.com/8HMBENUS/2731??PS=47575