Abstract

Let {Xi, 1 ≤ i ≤ n} be a random sample from a continuous dƒ F(x) with the λ-th quantile q(λ) = inf (x : F(x) > λ). The present work concerns with the study of the smoothed kernel quantile estimator q(λ) proposed by Yang (1985) and establish in terms of mean square error its asymptotic viz. LIL, Berry-Esseen’s, theorem. By Monte-Carlo study establish, its superiority over all other forms of quantile estimators of q(λ) considered so far in the literature.

Author: K. Padmavathi and Y. S. Rama Krishnaiah

Received on: May, 2009

Accepted on: December, 2009