Abstract
This work links the conditional probability structure of Lancaster probabilities to a construction of reversible continuous-time Markov processes. Such a task is achieved by using the spectral expansion of the corresponding transition probabilities in order to introduce a continuous time dependence in the orthogonal representation inherent to Lancaster probabilities. This relationship provides a novel methodology to build continuous-time Markov processes via Lancaster probabilities. Particular cases of well-known models are seen to fall within this approach. As a byproduct, it also unveils new identities associated to well known orthogonal polynomials.
Subject
Statistics and Probability
Reference26 articles.
1. Bateman H. and
Erdélyi A., Vol. 1 of
Higher transcendental functions.
McGraw-Hill,
New York
(1953).
2. Björk T.,
Arbitrage theory in continuous time.
Oxford University Press
(2009).
3. Gibbs Sampling, Exponential Families and Orthogonal Polynomials
4. Polynomial Expansions of Bivariate Distributions