Pascal processes and their characterization
Let ([Pi]t) be a counting process on + with the property that for any t, T with 0[less-than-or-equals, slant]t[less-than-or-equals, slant]T the distribution of [Pi]T given the past t is Pascal (negative binomial) with one parameter being [Pi]t+1 and the probability parameter depending only on t and T. Does such a process exist? If so, how is it characterized? Finally, what is the most convenient way to model such a process? These questions are motivated by the distinguished role of the Pascal distribution in finding explicit solutions of optimal selection problems based on relative ranks. We answer them completely.
| Year of publication: |
1991
|
|---|---|
| Authors: | Bruss, F. T. ; Rogers, L. C. G. |
| Published in: |
Stochastic Processes and their Applications. - Elsevier, ISSN 0304-4149. - Vol. 37.1991, 2, p. 331-338
|
| Publisher: |
Elsevier |
| Keywords: | Yule processes mixed Poisson processes record processes martingales |
Saved in:
Saved in favorites
Similar items by person
-
On the maximum and its uniqueness for geometric random samples
Bruss, F. T., (1990)
-
Rogers, Leonard C. G., (2008)
-
Sense, nonsense and the S&P500
Rogers, Leonard C. G., (2018)
- More ...