Geometric stable distributions in Banach spaces (Q1322912): Difference between revisions

The paper deals with the class of geometric stable (GS) distributions whose significance for heavy tailed modeling in various areas is illustrated by examples and documented by a large number of references. On a Banach space \(B\) a characterization of GS distributions as well as of the more general class of geometric infinitely divisible distributions is given in terms of characteristic functions and limits of sums of i.i.d. r.v.'s. For the case that \(Z_ p:= p^{1/\alpha} \sum_{i=1}^{T(p)} X_ i\) converges for \(p\to 0\) weakly towards a \(B\)-valued r.v. \(Y\) with GS law -- where \(X_ 1, X_ 2,\dots\) are i.i.d. \(B\)-valued r.v.'s, \(T(p)\) is a geometrically distributed r.v. independent of the sequence \(\{X_ i\}\), \(0<\alpha \leq 2\), \(p\in [0,1]\) -- the authors give convergence rates for \(\| Z_ p\| @>w >> \| Y\|\) and \(Z_ p @> w>> Y\) in terms of certain ideal metrics of various orders and the Prokhorov metric.