Probabilities of large deviations in Banach spaces

Probability theory on linear topological spaces (60B11) Limit theorems for vector-valued random variables (infinite-dimensional case) (60B12) Large deviations (60F10)

Review text

Let \(X_ j,j=1,2,...,n\) be independent random elements in a separable Banach space \({\mathcal E}\), \(S=\sum^{n}_{j=1}X_ j\), \(A_ t=\sum^{n}_{j=1}E \| X_ j\|^ t\), \(EX_ j=0\), \(j=1,2,...,n\). It is proved that if \(\alpha>1\) then \[ P[\| S\| \geq U]\geq [1-A_ 2u^{-2}(\frac{1}{\alpha^ 2}+\frac{1}{A_ 2u^{- 2}+(\alpha -1)^ 2})]\sum^{n}_{j=1}P[\| X_ j\| \geq \alpha u] \] for all \(u>0\), and if \(t>2\) then \[ E\| S\|^ t\geq \min [(\frac{E\| S\|^ 2}{A_ 2})^{t/2},2^{-5}t]\max [2^{2^{- t}}t^{-1}A_ t,A_ 2^{t/2}]. \] The last inequality is a generalization of an \textit{H. P. Rosenthal's} [Isr. J. Math. 8, 273-303 (1970; Zbl 0213.193)] analogous result.

Recommendations

Cites work

Cited in

(11)

This page was built for publication: Probabilities of large deviations in Banach spaces

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q791959)