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")