Limit distributions of Studentized means.

From MaRDI portal





Let \(\{X_n\}\) be a sequence of i.i.d. random variables, \(S_n= \sum^n_{j=1}X_j\) and \(V_n^2=\sum^n_{j=1}X^2_j\). The problem of limiting behaviour of self-normalized sums \(S_n/V_n\) is completely solved. It is proved that \(S_n/V_n\) converges weakly to a limit \(Z\) with \(P(| Z|=1)\neq 1\) if and only if \(X_1\) is in the domain of attraction of a stable law with some exponent \(0< \alpha\leq 2\) and \(X_1\) is centered when \(\alpha>1\) or Feller's condition holds when \(\alpha=1\). Moreover, \(S_n/V_n\) converges weakly to \(Z\) with \(P(| Z |=1) =1\) if and only if \(P(| X_1|>x)\) is a slowly varying function at infinity. No further assumptions on \(X_j\)'s, like e.g., symmetry, are made.



Cites work


Cited in
(44)








This page was built for publication: Limit distributions of Studentized means.

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1879845)