https://github.com/cran/dse
Tip revision: f6673198843cfdf7ab8542715823fa692f1eb83c authored by Paul Gilbert on 07 March 2013, 00:00:00 UTC
version 2013.3-2
version 2013.3-2
Tip revision: f667319
informationTestsCalculations.Rd
\name{informationTestsCalculations}
\alias{informationTestsCalculations}
\title{Calculate selection criteria}
\description{Calculates several model selection criteria.}
\usage{
informationTestsCalculations(lst, sample.start=1,sample.end=NULL, warn=TRUE)
}
\arguments{
\item{lst}{One or more objects of class TSestModel.}
\item{sample.start}{
The start of the period to use for criteria calculations.
}
\item{sample.end}{
The end of the period to use for criteria calculations.
If omitted the end of the sample is used.
}
\item{warn}{If FALSE then some warning messages are suppressed.}
}
\value{
The calculated values are returned in a vector with names: port, like,
aic, bic, gvc, rice, fpe, taic, tbic, tgvc, trice and tfpe. These
correspond to values for the Portmanteau test, likelihood,
Akaike Information Criterion, Bayes Information Criterion,
Generalized Cross Validation, Rice Criterion, and Final Prediction Error.
The preceeding 't' indicates that the theoretical parameter space
dimension has been used, rather than the number of
coefficient (parameter) values. Methods which select a model based on some
information criterion calculated by \code{informationTestsCalculations}
should use the name of the vector element to specify the test value which
is to be used.
}
\seealso{
\code{\link{informationTests}}
}
\examples{
data("eg1.DSE.data.diff", package="dse")
model <- estVARXls(eg1.DSE.data.diff)
informationTestsCalculations(model)
}
\concept{DSE}
\keyword{ts}