Multivariate order statistics via multivariate concomitants (Q1006674): Difference between revisions

Let \(\mathbf{X}_1,\mathbf{X}_2,\ldots,\mathbf{X}_n\) be independent and identically distributed \(k\)-dimensional random vectors. In an effort to define order statistics based on these random vectors, \textit{I. Bairamov} [ibid. 97, No. 4, 797-809 (2006; Zbl 1085.62062)] suggested to order the \(\mathbf{X}_i\)'s according to the ordered values of some real function of these random vectors. The present authors generalize this idea by introducing an additional dimension, and an additional random variable that is augmented to the corresponding random vectors, giving the independent and identically distributed \((k+1)\)-dimensional random vectors \((\mathbf{X}_1,Y_1),(\mathbf{X}_2,Y_2),\ldots,(\mathbf{X}_n,Y_n)\). They then define the \(j\)\,th smallest order statistic among the \(n\) \(\mathbf{X}_i\)'s as the concomitant of the \(j\)\,th smallest \(Y_i\). The distribution of the \(j\)\,th smallest multivariate order statistic is derived. Some applications and examples are given.