A robust algorithm for parameter estimation in smooth transition autoregressive models

被引:5
|
作者
Bekiros, Stelios D. [1 ]
机构
[1] European Univ Inst, I-50014 Fiesole, FI, Italy
关键词
STAR models; Gradient descent; Singular value decomposition; Heteroscedastic noise;
D O I
10.1016/j.econlet.2009.01.020
中图分类号
F [经济];
学科分类号
02 ;
摘要
Finding a precise estimate for the smoothness parameter of LSTAR models is notoriously difficult. This paper introduces a robust estimation method for the transition and autoregressive parameters of STAR models, comprising gradient descent and singular value decomposition to account for heteroscedastic noise. (C) 2009 Elsevier B.V. All rights reserved.
引用
收藏
页码:36 / 38
页数:3
相关论文
共 50 条