1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
|
\name{jack.test}
\alias{jack.test}
\alias{print.jacktest}
\title{Jackknife approximate t tests of regression coefficients}
\description{
Performes approximate t tests of regression coefficients
based on jackknife variance estimates.
}
\usage{
jack.test(object, ncomp = object$ncomp, use.mean = TRUE)
\method{print}{jacktest}(x, P.values = TRUE, \dots)
}
\arguments{
\item{object}{an \code{mvr} object. A cross-validated model fitted
with \code{jackknife = TRUE}.}
\item{ncomp}{the number of components to use for estimating the variances}
\item{use.mean}{logical. If \code{TRUE} (default), the mean
coefficients are used when estimating the (co)variances; otherwise
the coefficients from a model fitted to the entire data set. See
\code{\link{var.jack}} for details.}
\item{x}{an \code{jacktest} object, the result of \code{jack.test}.}
\item{P.values}{logical. Whether to print \eqn{p} values (default).}
\item{\dots}{Further arguments sent to the underlying print function
\code{\link{printCoefmat}}.}
}
\details{
\code{jack.test} uses the variance estimates from \code{var.jack} to
perform \eqn{t} tests of the regression coefficients. The resulting object
has a print method, \code{print.jacktest}, which uses
\code{\link{printCoefmat}} for the actual printing.
}
\value{
\code{jack.test} returns an object of class \code{"jacktest"}, with components
\item{coefficients }{The estimated regression coefficients}
\item{sd}{The square root of the jackknife variance estimates}
\item{tvalues}{The \eqn{t} statistics}
\item{df}{The `degrees of freedom' used for calculating \eqn{p}
values}
\item{pvalues}{The calculated \eqn{p} values}
\code{print.jacktest} returns the \code{"jacktest"} object (invisibly).
}
\section{Warning}{
The jackknife variance estimates are known to be biased (see
\code{\link{var.jack}}).
Also, the distribution of the regression coefficient estimates and the
jackknife variance estimates are unknown (at least in PLSR/PCR).
Consequently, the distribution (and in particular, the degrees of
freedom) of the resulting \eqn{t} statistics is unknown. The present code
simply assumes a \eqn{t} distribution with \eqn{m - 1} degrees of
freedom, where \eqn{m} is the number of cross-validation segments.
Therefore, the resulting \eqn{p} values should not be used
uncritically, and should perhaps be regarded as mere indicator of
(non-)significance.
Finally, also keep in mind that as the number of predictor variables
increase, the problem of multiple tests increases correspondingly.
}
\references{
Martens H. and Martens M. (2000) Modified Jack-knife Estimation of
Parameter Uncertainty in Bilinear Modelling by Partial Least Squares
Regression (PLSR). \emph{Food Quality and Preference}, \bold{11}, 5--16.
}
\author{Bjørn-Helge Mevik}
\seealso{\code{\link{var.jack}}, \code{\link{mvrCv}}}
\examples{
data(oliveoil)
mod <- pcr(sensory ~ chemical, data = oliveoil, validation = "LOO", jackknife = TRUE)
jack.test(mod, ncomp = 2)
}
\keyword{htest}
|
