ASYMPTOTIC RESULTS FOR GOODNESS-OF-FIT STATISTICS WHEN PARAMETERS MUST BE ESTIMATED.
STANFORD UNIV CALIF DEPT OF STATISTICS
Pagination or Media Count:
The paper considers the asymptotic distribution of the goodness-of-fit statistics W squared and U squared when parameters of the tested distribution must be estimated from the sample. The asymptotic mean and variance are found for the cases when, in testing for normality, either mu or sigma squared or both must be estimated, and when, in testing for exponentiality, the scale parameter must be estimated. Section 2 also gives a list of integrals connected with the normal distribution, which might have other uses. Author
- Statistics and Probability