https://github.com/cran/MuMIn
Tip revision: 6c59be08a9eaa84f2923153255821df7509a857e authored by Kamil Bartoń on 02 October 2017, 11:39:13 UTC
version 1.40.0
version 1.40.0
Tip revision: 6c59be0
subset.model.selection.Rd
\name{subset.model.selection}
\alias{subset.model.selection}
\alias{[.model.selection}
\alias{[[.model.selection}
\alias{has}
\newcommand{\bq}{\verb{`}\code{#1}\verb{`}}
\encoding{utf-8}
\title{Subsetting model selection table}
\description{Extract subset of a model selection table.
}
\usage{
\method{subset}{model.selection}(x, subset, select, recalc.weights = TRUE, recalc.delta = FALSE, ...)
\method{[}{model.selection}(x, i, j, recalc.weights = TRUE, recalc.delta = FALSE, ...)
\method{[[}{model.selection}(x, ..., exact = TRUE)
}
\arguments{
\item{x}{a \code{model.selection} object to be subsetted.}
\item{subset,select}{logical expressions indicating columns and rows to keep.
See \code{\link{subset}}. }
\item{i,j}{indices specifying elements to extract. }
\item{recalc.weights}{logical value specyfying whether Akaike weights
should be normalized across the new set of models to sum to one.
}
\item{recalc.delta}{logical value specyfying whether
\ifelse{latex}{\eqn{\Delta_{IC}}}{\enc{Δ}{Delta}_IC}
should be calculated for the new set of models (not done by default).
}
\item{exact}{logical, see \code{\link[=Extract]{[}}}.
\item{\dots}{further arguments passed to \code{\link{[.data.frame}}
(\code{drop}). }
}
\value{
A \code{model.selection} object containing only the selected models (rows).
If columns are selected (via argument \code{select} or the second index
\code{x[, j]}) and not all essential columns (i.e. all except
"varying" and "extra") are present in the result, a plain \code{data.frame} is
returned. Similarly, modifying values in the essential columns with \code{[<-},
\code{[[<-} or \code{$<-} produces a regular data frame.
}
\details{
Unlike the method for \code{data.frame}, single bracket extraction with only
one index \code{x[i]} selects rows (models) rather than columns.
To select rows according to presence or absence of the variables (rather than
their value), a pseudo-function \code{has} may be used with \code{subset}, e.g.
\code{subset(x, has(a, !b))}
will select rows with \emph{a} \bold{and} without \emph{b} (this is
equivalent to \code{!is.na(a) & is.na(b)}). \code{has} can take any number of
arguments.
Complex model terms need to be enclosed within curly brackets
(e.g \code{{s(a,k=2)}}), except for within \code{has}. Backticks-quoting is
also possible, but then the name must match exactly (including whitespace)
the term name as returned by \code{getAllTerms}.
Enclosing in \code{I} prevents a name from being interpreted as column name.
To select rows where one variable can be present conditional on the presence of
other variable(s), the function \code{dc} (\bold{d}ependency \bold{c}hain) can
be used.
\code{dc} takes any number of variables as arguments, and allows a variable to be
included only if all the preceding arguments are also included (e.g. \code{subset =
dc(a, b, c)} allows for models of form \code{a}, \code{a+b} and \code{a+b+c} but not
\code{b}, \code{c}, \code{b+c} or \code{a+c}).
}
\author{Kamil Barto\enc{ń}{n}}
\seealso{
\code{\link{dredge}}, \code{\link{subset}} and \code{\link{[.data.frame}} for
subsetting and extracting from \code{data.frame}s.
}
\examples{
fm1 <- lm(formula = y ~ X1 + X2 + X3 + X4, data = Cement, na.action = na.fail)
# generate models where each variable is included only if the previous
# are included too, e.g. X2 only if X1 is there, and X3 only if X2 and X1
dredge(fm1, subset = dc(X1, X2, X3, X4))
# which is equivalent to
# dredge(fm1, subset = (!X2 | X1) & (!X3 | X2) & (!X4 | X3))
# alternatively, generate "all possible" combinations
ms0 <- dredge(fm1)
# ...and afterwards select the subset of models
subset(ms0, dc(X1, X2, X3, X4))
# which is equivalent to
# subset(ms0, (has(!X2) | has(X1)) & (has(!X3) | has(X2)) & (has(!X4) | has(X3)))
# Different ways of finding a confidence set of models:
# delta(AIC) cutoff
subset(ms0, delta <= 4, recalc.weights = FALSE)
# cumulative sum of Akaike weights
subset(ms0, cumsum(weight) <= .95, recalc.weights = FALSE)
# relative likelihood
subset(ms0, (weight / weight[1]) > (1/8), recalc.weights = FALSE)
}
\keyword{manip}