Full text
| |
| Author(s): |
Alexsandro Giacomo Grimbert Gallo
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | São Paulo. |
| Institution: | Universidade de São Paulo (USP). Instituto de Matemática e Estatística (IME/SBI) |
| Defense date: | 2009-10-30 |
| Examining board members: |
Jefferson Antonio Galves;
Nancy Lopes Garcia;
Aurelien Garivier;
Florencia Graciela Leonardi;
Jorma Johannes Rissanen
|
| Advisor: | Jefferson Antonio Galves |
| Field of knowledge: | Physical Sciences and Mathematics - Probability and Statistics |
| Indexed in: | Banco de Dados Bibliográficos da USP-DEDALUS |
| Location: | Universidade de São Paulo. Instituto de Matemática e Estatística. Biblioteca Carlos Benjamin de Lyra; IME-T QA274.8.T; G172s |
| Abstract | |
We present a new perfect simulation algorithm for stationary chains (indexed by Z) having unbounded variable length memory. This is the class of innite memory chains for which the family of transition probabilities is given by probabilistic context tree. Our condition is expressed in terms of the structure of the context tree. In particular, we do not assume the continuity of the family of transition probabilities. We give an explicit construction of the chain using a sequence of i.i.d. random variables uniformly distributed in [0,1[. (AU) | |