Inference on a distribution function from ranked set samples (Q2304241)

Consider a sample of \(k\) independent observations \((X_i,R_i)\), with random or fixed rank \(R_i\), \(i=1,2,\dots,k\). Here, the random variable \(X_i\), given rank \(R_i=r\), has the conditional probability distribution \(F_r(x) = B_r(F(x))\), where \(B_r\) denotes the distribution function of the beta distribution with parameters \(r, k+1-r\), and \(F\) is an unknown distribution function to be estimated. The properties of three different estimators of \(F=F(x)\), a stratified estimator, a nonparametric maximum likelihood estimator and a moment based estimator, are considered. Especially, a functional central limit theorem is given for evaluating the asymptotic behavior, and pointwise and simultaneous confidence intervals for the distribution function with guaranteed coverage probability for finite samples. The estimation procedures are illustrated by means of a real data example.