Random Weight ting the Functional Statistics

Abstract:

      Let Tn be a functional of sample empirical distribution function T_n = T(F_n). With some conditions,n^1/2(T( F_n) - T (F)) is asymptotic normality. In this paper, we use the ideal of random weightting to approximate the distribution of n^1/2.(T(F_n) -T(F)).For Von-Mises statistics, U - statistics and L - statistics,the detail disussion has been made.