a numeric vector of weights to be given to individual observations
in the sum of squared residuals that forms the local fitting criterion.
robust
a numeric vector of robustness weights
span
smoothing parameter.
degree
overall degree of locally-fitted polynomial.
1 is locally-linear fitting and
2 is locally-quadratic fitting.
parametric
for two or more numeric predictors, this argument
specifies those variables that should
be conditionally-parametric. It should be specified as
a logical vector of length equal to the number of columns in
x.evaluate.
drop.square
for cases with
degree equal to
2 and with two or more numeric predictors,
this argument
specifies those numeric predictors whose squares should be dropped from
the set of fitting
variables.
The method of specification is the same as
for
parametric.
surface
determines whether the fitted surface is computed directly
at all points (
"direct") or whether an interpolation method is
used (
"interpolate").
statistics
determines whether the statistical quantities are computed exactly
(
"exact") or approximately (
"approximate") (
"none" implies no
statistical calculations).
cell
the maximum cell size of the k-d tree.
Suppose k <- floor(n*cell) where
n is the number of observations.
Then a cell is further divided if the number of observations within it
is greater than or equal to
k.
trace.hat
determines the computational method used to compute the trace of the hat
matrix, which is used in the computation of the statistical quantities.
(
"exact" or
"approximate", please refer to documentation of
loess.control for usage detail).
VALUE:
This is a support routine for
loess. It returns the fitted values,
the k-d tree if interpolations is used, and the statistics if
statistics
is not
"none".