Stochastic Inequalities Relating a Class of Log-Likelihood Ratio Statistics to their Asymptotic $\chi^2$ Distribution
Porteous, B. T.
Ann. Statist., Tome 17 (1989) no. 1, p. 1723-1734 / Harvested from Project Euclid
For decomposable covariance selection models, stochastic inequalities which relate the null distribution of the log-likelihood ratio statistic to its asymptotic $\chi^2$ distribution are obtained. The implications are twofold: First, the null distribution of the log-likelihood ratio statistic is seen to be stochastically larger than its asymptotic $\chi^2$ distribution. Extremely large samples apart, for the $\chi^2$ approximation to be valid, a deflation of the log-likelihood ratio statistic is then necessary. Second, a simple adjustment to the log-likelihood ratio statistic, similar in spirit to the Bartlett adjustment, yields a conservative test.
Publié le : 1989-12-14
Classification:  Asymptotic distribution,  Bartlett adjustment,  conservative test,  covariance selection,  decomposability,  multivariate analysis,  partitioning,  62E15,  62H99
@article{1176347390,
     author = {Porteous, B. T.},
     title = {Stochastic Inequalities Relating a Class of Log-Likelihood Ratio Statistics to their Asymptotic $\chi^2$ Distribution},
     journal = {Ann. Statist.},
     volume = {17},
     number = {1},
     year = {1989},
     pages = { 1723-1734},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1176347390}
}
Porteous, B. T. Stochastic Inequalities Relating a Class of Log-Likelihood Ratio Statistics to their Asymptotic $\chi^2$ Distribution. Ann. Statist., Tome 17 (1989) no. 1, pp.  1723-1734. http://gdmltest.u-ga.fr/item/1176347390/