Page 1 of 1

Likelihood Ratio Test

Posted: Sun Jul 25, 2010 1:37 pm
by 00tailor
Hi,

as I have to use LR Test (unconditional coverage) by Kupiec (1995), but unfortunately do have no experience with log likelihood estimation, please could anybody write a code or give some hints for a program for this LR Test:

LR= 2{log[(f^x)*((1-f)^(T-x))]-log[(alpha^x)*((1-alpha)^(T-x))]} ~ chi²,

where f and alpha are scalar and x is a number of observations, which is a part of total observations T.

Thank you in advance.