Integer valued Markov processes and exponential families
We characterize all continuous-time Markov Processes on the set of integers such that their distribution at any instant is in a natural exponential family with one parameter. A complete description of birth and death processes having this property is given. They include linear growth birth and death processes as well as new examples which are explicitly treated. AMS 1980 Subject Classifications: 60J25, 62E10
Year of publication: |
1992
|
---|---|
Authors: | Ycart, Bernard |
Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 14.1992, 1, p. 71-78
|
Publisher: |
Elsevier |
Keywords: | Markov processes exponential families birth and death processes |
Saved in:
Saved in favorites
Similar items by person
-
Kleptsyna, Marina, (2014)
-
Zero-one Laws for Binary Random Fields
Coupier, David, (2005)
-
Cut-off for n-tuples of exponentially converging processes
Barrera, Javiera, (2006)
- More ...