Πολυμεταβλητές μεμειγμένες διαδικασίες Poisson
Multivariate mixed Poison processes
View/ Open
Subject
Poisson processesKeywords
Πολυμεταβλητές στατιστικές μέθοδοιAbstract
In this thesis we study the multivariate mixed Poisson processes with arbitrary
mixing distribution and the multivariate mixed Poisson processes with parameter
a random vector. Some properties of multivariate mixed Poisson processes, such
as the multinomial and the Markov property, are derived. The use of multivariate
setting is justified by a result, which asserts that the coordinates of a multivariate
mixed Poisson process are independent, if and only if the mixing distribution is
represented as a product measure. Moreover some characterizations for multivariate
mixed Poisson processes, in terms of the multinomial and the Markov property are
given. Finally, it is proven that the class of all multivariate mixed Poisson processes
with parameter a random vector is subclass of the class of all multivariate mixed
Poisson processes with a mixing distribution. The problem of the equality of the
two classes remains open.