https://github.com/cran/dse
Tip revision: 351baacd22b7252d0d358ff1d9c2c11e14ab0c81 authored by Paul Gilbert on 14 October 2009, 00:00:00 UTC
version 2009.10-2
version 2009.10-2
Tip revision: 351baac
residualStats.Rd
\name{residualStats}
\alias{residualStats}
\title{Calculate Residuals Statistics and Likelihood}
\description{Calculate the residuals statistics and likelihood of a residual.}
\usage{
residualStats(pred, data, sampleT=nrow(pred), warn=TRUE)}
\arguments{
\item{pred}{A matrix with columns representing time series.}
\item{data}{A matrix with columns representing time series.}
\item{sampleT}{An integer indicating the sample to use.}
\item{warn}{If FALSE certain warnings are suppressed.}
}
\details{
Residuals are calculated as
pred[1:sampleT,,drop=FALSE] - data[1:sampleT,,drop=FALSE]
and then statistics are calculated based on these residuals. If pred
or data are NULL they are treated as zero.
}
\value{
A list with elements like, cov, pred, and sampleT. pred and sampleT are
as supplied (and are returned as this is a utility function called by
other functions and it is convenient to pass them along). cov is the
covariance of the residual and like is a vector of four elements
representing the total, constant, determinant and covariance
terms of the negative log likelihood function.
}
\seealso{
\code{\link{l}}
}
\examples{
residualStats(matrix(rnorm(200), 100,2), NULL) # but typically used for a residual
}
\concept{DSE}
\keyword{ts}