The probabilistic structure of a discrete-time (high-order) vector Markov process may be studied using two approaches. In the first approach, the Markov process is specified by the transition probability and the initial distribution. An alternative approach is via a stochastic difference equation. We have proved that these two approaches are equivalent under very mild conditions.
Publié le : 2002-02-15
Classification:
ARMA model,
equivalence of distribution functions,
stationarity,
stochastic difference equation,
transition probability
@article{1078951092,
author = {Chan, K.-S. and Tong, H.},
title = {A note on the equivalence of two approaches for specifying a Markov process},
journal = {Bernoulli},
volume = {8},
number = {2},
year = {2002},
pages = { 117-122},
language = {en},
url = {http://dml.mathdoc.fr/item/1078951092}
}
Chan, K.-S.; Tong, H. A note on the equivalence of two approaches for specifying a Markov process. Bernoulli, Tome 8 (2002) no. 2, pp. 117-122. http://gdmltest.u-ga.fr/item/1078951092/