Maximum likelihood estimation for an observation driven model for Poisson counts

被引：28

作者：

Davis, RA ^{[1
]}

Dunsmuir, WTM

Streett, SB

机构：

[1] Colorado State Univ, Dept Stat, Ft Collins, CO 80523 USA

[2] Univ New S Wales, Dept Stat, Sydney, NSW, Australia

[3] Natl Inst Stand & Technol, Stat Engn Div, Boulder, CO USA

来源：

METHODOLOGY AND COMPUTING IN APPLIED PROBABILITY | 2005年 / 7卷 / 02期

基金：

美国国家科学基金会;

关键词：

observation-driven model; Poisson valued time series;

D O I：

10.1007/s11009-005-1480-4

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 [统计学]; 070103 [概率论与数理统计]; 0714 [统计学];

摘要：

This paper is concerned with an observation-driven model for time series of counts whose conditional distribution given past observations follows a Poisson distribution.This class of models is capable of modeling a wide range of dependence structures and is readily estimated using an approximation to the likelihood function. Recursive formulae for carrying out maximum likelihood estimation are provided and the technical components required for establishing a central limit theorem of the maximum likelihood estimates are given in a special case.

引用

页码：149 / 159

页数：11