Non-parametric inferences based on general unbalanced ranked-set samples

Abstract

A general unbalanced ranked-set sample consists of independent order statistics each of which is out of a subsample from a common population. Such data can arise from two situations: (a) a designed ranked-set sampling (RSS) and (b) certain experimental process, e.g., the r-out-of-k systems in life testing experiments. There is no well accepted approach available so far in the literature for the effective analysis of such data. In this article, we develop methods for making inferences on various features of the population such as quantile, distribution function and moments etc., based on data of the above nature. The asymptotic properties of the methods are well established. Some simulation results are also provided for the vindication of the methods.