This package provides modified versions of
contr.treatment
and
contr.sum
. contr.sum
gains an optional base
argument, analog to the
one of contr.treatment
, furthermore,
the base
argument may be the name of a
factor level.
contr
returns a function that calls either
contr.treatment
, contr.sum
, etc.,
according to the value given to its first argument.
The contrasts
method for "item"
objects
returns a contrast matrix or a function to produce
a contrast matrix for the factor into which
the item would be coerced via as.factor
or as.ordered
.
This matrix or function can be specified by
using contrasts(x)<-value
contr(type,…)
contr.treatment(n, base=1,contrasts=TRUE)
contr.sum(n,base=NULL,contrasts=TRUE)
# S4 method for item
contrasts(x,contrasts=TRUE,…)
# S4 method for item
contrasts(x,how.many) <- value
# These methods are defined implicitely by making 'contrasts' generic.
# S4 method for ANY
contrasts(x,contrasts=TRUE,…)
# S4 method for ANY
contrasts(x,how.many) <- value
a character vector, specifying the type of the contrasts.
This argument should have a value such that, if e.g. type="something"
,
then there is a function contr.something
that produces
a contrast matrix.
further arguments, passed to contr.treatment
, etc.
a number of factor levels or a vector of factor levels names, see e.g. contr.treatment
.
a number of a factor level or the names of a factor level,
which specifies the baseline category,
see e.g. contr.treatment
or NULL.
a logical value, see contrasts
the number of contrasts to generate, see contrasts
a factor or an object of class "item"
a matrix, a function or the name of a function
contr
returns a funtion that calls one of contr.treatment
,
contr.sum,…
.
contr.treatment
and contr.sum
return contrast matrices.
contrasts(x)
returns the "contrasts" attribute of an
object, which may be a function name, a function, a contrast matrix or NULL.
# NOT RUN {
ctr.t <- contr("treatment",base="c")
ctr.t
ctr.s <- contr("sum",base="c")
ctr.h <- contr("helmert")
ctr.t(letters[1:7])
ctr.s(letters[1:7])
ctr.h(letters[1:7])
x <- factor(rep(letters[1:5],3))
contrasts(x)
x <- as.item(x)
contrasts(x)
contrasts(x) <- contr.sum(letters[1:5],base="c")
contrasts(x)
missing.values(x) <- 5
contrasts(x)
contrasts(as.factor(x))
# Obviously setting missing values after specifying
# contrast matrix breaks the contrasts.
# Using the 'contr' function, however, prevents this:
missing.values(x) <- NULL
contrasts(x) <- contr("sum",base="c")
contrasts(x)
missing.values(x) <- 5
contrasts(x)
contrasts(as.factor(x))
# }
Run the code above in your browser using DataLab