These are predefined prepanel functions available in Lattice.
prepanel.lmline(x, y, …)
prepanel.qqmathline(x, y = x, distribution = qnorm,
probs = c(0.25, 0.75), qtype = 7,
groups, subscripts,
…)
prepanel.loess(x, y, span, degree, family, evaluation,
horizontal = FALSE, …)
prepanel.spline(x, y, npoints = 101,
horizontal = FALSE, …,
keep.data = FALSE)x and y values, numeric or factor
quantile function for theoretical
distribution. This is automatically passed in when this is used as a
prepanel function in qqmath.
type of quantile
numeric vector of length two, representing probabilities. If used
with aspect="xy", the aspect ratio will be chosen to make the
line passing through the corresponding quantile pairs as close to 45
degrees as possible.
Arguments controlling the
underlying loess smooth.
See documentation for corresponding panel function.
Ignored. Present to capture argument of the same
name in smooth.spline.
See xyplot. Whenever
appropriate, calculations are done separately for each group and
then combined.
Other arguments. These are passed on to other functions
if appropriate (in particular, smooth.spline), and
ignored otherwise.
usually a list with components xlim, ylim, dx and
dy, the first two being used to calculate panel axes limits,
the last two for banking computations. The form of these components
are described under xyplot. There are also several
prepanel functions that serve as the default for high level functions,
see prepanel.default.xyplot
All these prepanel functions compute the limits to be large enough to contain all points as well as the relevant smooth.
In addition, prepanel.lmline computes the dx and
dy such that it reflects the slope of the linear regression
line; for prepanel.qqmathline, this is the slope of the line
passing through the quantile pairs specified by probs. For
prepanel.loess and prepanel.spline, dx and
dy reflect the piecewise slopes of the nonlinear smooth.