scientific article; zbMATH DE number 3868334

From MaRDI portal
Publication:3334713





[For the entire collection see Zbl 0533.00016.] Let \(X_{in}\), \(i=\overline{1,N_ n}\), \(n\geq 1\), be a series of sequences of random variables (r.v.) with \(EX_{in}=0\), \(S_ N=\sum^{N}_{j=1}X_{jn}\), \(B^ 2_ N=ES^ 2_ N\). Let for the events \(A\in \sigma \{X_{1n},...,X_{tn}\}\) and \(B\in \sigma \{X_{t+j,n},...,X_{Nn}\}\) the condition \(\sup_ t| P(AB)- P(A)P(B)| \leq \phi(j)\) be satisfied. Assume that \(\phi\) (j)\(\leq 1\), \(\phi\) (j)\(\downarrow 0\), \(j\to \infty\) and \(\sum^{\infty}_{j=0}\phi^{{1\over2}}(j)<\infty\). The asymptotic behaviour of the probability \(P\{S_ N/B_ N\geq(\ln B_ N)^{{1\over2}}\}\) under restrictions on the moments of \(X_{in}\) and variation \(B_ N\) is investigated. The theorem proved generalizes the result of \textit{M. Ghosh} and \textit{G. J. Babu} [Ann. Probab. 5, 222-234 (1977; Zbl 0382.60036)].











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3334713)