uTAR.est: General Estimation of TAR Models

In NTS: Nonlinear Time Series Analysis

General Estimation of TAR Models

Description

General estimation of TAR models with known threshold values. It perform LS estimation of a univariate TAR model, and can handle multiple regimes.

Usage

uTAR.est(
  y,
  arorder = c(1, 1),
  thr = c(0),
  d = 1,
  thrV = NULL,
  include.mean = c(TRUE, TRUE),
  output = TRUE
)

Arguments

y	time series.
arorder	AR order of each regime. The number of regime is the length of arorder.
thr	given threshold(s). There are k-1 threshold for a k-regime model.
d	delay for threshold variable, default is 1.
thrV	external threshold variable if any. If it is not NULL, thrV must have the same length as that of y.
include.mean	a logical value indicating whether constant terms are included. Default is TRUE.
output	a logical value for output. Default is TRUE.

Value

uTAR.est returns a list with components:

data	the data matrix, y.
k	the number of regimes.
arorder	AR orders of regimes 1 and 2.
coefs	a k-by-(p+1) matrices, where k is the number of regimes. The i-th row shows the estimation results in regime i.
sigma	estimated innovational covariances for all the regimes.
thr	threshold value.
residuals	estimated innovations.
sresi	standardized residuals.
nobs	numbers of observations in different regimes.
delay	delay for threshold variable.
cnst	logical values indicating whether the constant terms are included in different regimes.
AIC	AIC value.

Examples

phi=t(matrix(c(-0.3, 0.5,0.6,-0.3),2,2))
y=uTAR.sim(nob=200, arorder=c(2,2), phi=phi, d=2, thr=0.2, cnst=c(1,-1),sigma=c(1, 1))
thr.est=uTAR(y=y$series, p1=2, p2=2, d=2, thrQ=c(0,1),Trim=c(0.1,0.9), method="RLS")
est=uTAR.est(y=y$series, arorder=c(2,2), thr=thr.est$thr, d=2)

NTS documentation built on Sept. 25, 2023, 1:08 a.m.