On path properties of certain infinitely divisible processes
Let {X(t): t [set membership, variant] T} be a stochastic process equal in distribution to {[integral operator]sf(t, s)[Lambda](ds): t [set membership, variant] T}, where [Lambda]is a symmetric independently scattered random measure and f is a suitable deterministic function. It is shown that various properties of the sections f(·,s), s [set membership, variant] S, are inherited by the sample paths of X, provided X has no Gaussian component. The analogous statement for Gaussian processes is false. As a main tool, LePage-type series representation is fully developed for symmetric stochastic integral processes and this may be of independent interest.
Year of publication: |
1989
|
---|---|
Authors: | Rosinski, Jan |
Published in: |
Stochastic Processes and their Applications. - Elsevier, ISSN 0304-4149. - Vol. 33.1989, 1, p. 73-87
|
Publisher: |
Elsevier |
Keywords: | infinitely divisible processes sample path properties series and stochastic integral representations of infinitely divisible processes |
Saved in:
Saved in favorites
Similar items by person
-
Asymptotic bounds for infinitely divisible sequences
Kwapien, Stanislaw, (2006)
-
Modeling and simulation with operator scaling
Cohen, Serge, (2010)
-
On the oscillation of infinitely divisible and some other processes
Cambanis, Stamatis, (1990)
- More ...