plot.preplot.loess(object, xlab = object$xlab, ylab = object$ylab, which.plots = object$which.plots, rows, columns, show.given = TRUE, ...)
preplot.loess()
when the user wishes to save the evaluations of
the surface that are carried out to make the coplot.
formula
; if not,
the elements are assigned, in order, according to the order
in which the predictors appear in
formula
.
k
be the number of given values; if
columns
is missing,
then
rows <- ceiling(sqrt(k))
, else
rows <- ceiling(k/columns)
.
This argument is not used if there are two given predictors.
k
be the number of given values; if
rows
is missing,
columns <- ceiling(k/ceiling(sqrt(k)))
, else
columns <- ceiling(k/rows)
.
This argument is not used if there are two given predictors.
FALSE
, given panels are not included.
par
).
This function is a method for the generic function
for class preplot.loess.
It can be invoked by calling
for an object x of the appropriate class, or directly by calling
regardless of the class of the object.
This function graphs the fitted surface
of a local regression model for one, two or three predictors.
For one predictor, a curve is graphed against the predictor.
For two or three predictors, a coplot is made against each predictor,
conditional on the others. Each dependence panel of a coplot
shows a curve that is a slice through the surface and
is based on an evaluation for
evaluaton
equally-spaced values
of the predictor ranging between values specified by
ranges
;
in addition, confidence intervals at
confidence
equally-spaced
values over the same range are shown. These arguments are specified
during the creation of the preplot object by
preplot.loess()
.
Normally, the user will want to make
all coplots, but coplots against just certain predictors
can be made by using the argument
which.plots
.
air.m <- loess(ozone ~ radiation, data = air) air.m.pp <- preplot(air.m, given = 5, confidence = 7) plot(air.m.pp, which.plots = "radiation")