gets (version 0.9)

coef.gets: Extraction functions for 'gets' objects

Description

Extraction functions for objects of class 'gets'

Usage

"coef"(object, spec=NULL, ...) "fitted"(object, spec=NULL, ...) "logLik"(object, ...) "plot"(x, spec=NULL, col=c("red","blue"), lty=c("solid","solid"), lwd=c(1,1), ...) "predict"(object, spec=NULL, n.ahead=12, newmxreg=NULL, newvxreg=NULL, newindex=NULL, n.sim=1000, innov=NULL, return=TRUE, plot=TRUE, plot.options=list(), ...) "print"(x, ...) "residuals"(object, std=NULL, ...) "summary"(object, ...) "vcov"(object, spec=NULL, ...)

Arguments

object

an object of class 'gets'

x

an object of class 'gets'

spec

NULL, "mean", "variance" or, in some instances, "both". When NULL is a valid value, then it is automatically determined whether information pertaining to the mean or variance specification should be returned

std

logical. If FALSE (default), then the mean residuals are returned. If TRUE, then the standardised residuals are returned

n.ahead

generate forecasts up to n steps ahead (the default is 12)

newmxreg

a matrix (n.ahead rows and NCOL(mxregs) columns) with the out-of-sample values of the mxreg regressors

newvxreg

a matrix (n.ahead rows and NCOL(vxregs) columns) with the out-of-sample values of the vxreg regressors

newindex

date-index for the zoo object returned by predict.arx

n.sim

integer, the number of bootstrap replications for the generation of the variance forecasts

innov

NULL (default) or a vector of length n.ahead * n.sim containing the standardised errors (i.e. zero mean, unit variance) to bootstrap from

return

logical. If TRUE (default), then the out-of-sample forecasts are returned

plot

logical. If TRUE (default), then the out-of-sample forecasts are plotted

plot.options

a list of options related to the plotting of forecasts, see 'Details'

col

colours of fitted (default=red) and actual (default=blue) lines

lty

types of fitted (default=solid) and actual (default=solid) lines

lwd

widths of fitted (default=1) and actual (default=1) lines

...

additional arguments

Value

Details

The plot.options argument is a list that can contain any of the following arguments:

keep: integer greater than zero (default is 10) that controls the number of in-sample actual values to plot fitted: If TRUE, then the fitted values as well as actual values are plotted in-sample. The default is FALSE errors.only: logical or NULL (the default). If TRUE, then only mean forecasts are plotted when spec is set to "both" legend.loc: character string (the default is "topleft"). Allows the location of the plot legend to be altered newmactual: numeric vector or NULL (default). Enables the plotting of actual values out-of-sample in addition to the forecasts

See Also

getsm, getsv, isat

Examples

##Simulate from an AR(1):
set.seed(123)
y <- arima.sim(list(ar=0.4), 100)

##Simulate four independent Gaussian regressors:
xregs <- matrix(rnorm(4*100), 100, 4)

##estimate an AR(2) with intercept and four conditioning
##regressors in the mean, and a log-ARCH(3) in the variance:
mymod <- arx(y, mc=TRUE, ar=1:2, mxreg=xregs, arch=1:3)

##General-to-Specific (GETS) model selection of the mean:
meanmod <- getsm(mymod)

##General-to-Specific (GETS) model selection of the variance:
varmod <- getsv(mymod)

##print results:
print(meanmod)
print(varmod)

##plot the fitted vs. actual values, and the residuals:
plot(meanmod)
plot(varmod)

##print the entries of object 'gets':
summary(meanmod)
summary(varmod)

##extract coefficients of the simplified (specific) model:
coef(meanmod) #mean spec
coef(varmod) #variance spec

##extract log-likelihood:
logLik(mymod)

##extract variance-covariance matrix of simplified
##(specific) model:
vcov(meanmod) #mean spec
vcov(varmod) #variance spec

##extract and plot the fitted values:
mfit <- fitted(meanmod) #mean fit
plot(mfit)
vfit <- fitted(varmod) #variance fit
plot(vfit)

##extract and plot residuals:
epshat <- residuals(meanmod)
plot(epshat)

##extract and plot standardised residuals:
zhat <- residuals(varmod)
plot(zhat)