Nonparametric Confidence Intervals and Pitman Closeness of Joint Records to Population Quantiles
摘要
In this paper, based on the joint records from two or more independent sequences of continuous random variables, we develop outer and inner confidence intervals for quantile intervals, and upper and lower confidence limits for quantile differences. The coverage probabilities of these intervals are computed exactly. Furthermore, we study the Pitman closeness probabilities of upper joint records and derive various exact expressions. We also present the corresponding results for the special case of the power function distribution. A real data example illustrates the application of the proposed method