GENERALIZED CONFIDENCE INTERVALS FOR THE LARGEST VALUE OF SOME FUNCTIONS OF PARAMETERS UNDER NORMALITY

摘要: This paper deals with generalized confidence intervals (GCIs) for the maximum value of functions parameters interest in presence nuisance parameters. For k(≥ 2) normal populations, we propose GCIs for, respectively, largest mean, quantile and signal-to-noise ratio. case it is shown that proposed are better than those Chen Dudewicz (1973a, b). A new measure efficiency some Monte Carlo comparisons between method known performed. We also show several situations equivalent to Bayesian by employing improper prior distributions. Illustration made real data.