A general context-tree-based approach to perfect simulation for chains of infinite order
Preprint
Abstract
We establish new sufficient conditions for perfect simulation of chains of infinite order on a countable alphabet. This is done through the introduction of the \emph{localized continuity} assumption. This new assumption, which is formalized with the help of the notion of \emph{context trees}, includes the traditional continuous case, probabilistic context trees and discontinuous kernels. Since our assumptions are more refined than uniform continuity, our algorithms perfectly simulate continuous chains faster than the existing algorithms of the literature. We also give a simple extension of the method allowing to perfectly simulate chains which are neither necessarily continuous nor necessarily weakly non-null.Keywords
All Related Versions
This publication has 16 references indexed in Scilit:
- Context Trees, Variable Length Markov Chains and Dynamical Sources2012
- On Rate of Convergence of Statistical Estimation of Stationary Ergodic ProcessesIEEE Transactions on Information Theory, 2010
- Redundancy of the Context-Tree Weighting Method on Renewal and Markov Renewal ProcessesIEEE Transactions on Information Theory, 2006
- Bad configurations for random walk in random scenery and related subshiftsStochastic Processes and their Applications, 2005
- Intermittency and weak Gibbs statesNonlinearity, 2000
- Almost Gibbsian versus weakly Gibbsian measuresStochastic Processes and their Applications, 1999
- Perfect simulation and backward couplingCommunications in Statistics. Stochastic Models, 1998
- Redundancy rates for renewal and other processesIEEE Transactions on Information Theory, 1996
- Imbedded construction of stationary sequences and point processes with a random memoryQueueing Systems, 1994
- Random markov processes and uniform martingalesIsrael Journal of Mathematics, 1990