On least-squares and naive extrapolations in a non-linear AR(1) process

被引：12

作者：

Andel, J

机构：

[1] Charles University,Department of Statistics

来源：

TEST | 1997年 / 6卷 / 01期

关键词：

geometrically ergodic process; least squares extrapolation; naive extrapolation; non-linear autoregressive process;

D O I：

10.1007/BF02564427

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

A non-linear AR(1) process is investigated when the associated white noise has a rectangular distribution. The process is a modification of the logistic model and an important feature is that it is possible to derive explicit formulae for extrapolation. Some properties of the extrapolation are derived and it is proved that the least squares extrapolation m steps ahead converges to a constant as m --> infinity. The least squares extrapolation is compared with the naive extrapolation and the differences between them are shown to be small in some examples.

引用

页码：91 / 100

页数：10