Two dimensional local weighted least squares smoother. Only local linear smoother for estimating the original curve is available (no higher order, no derivative).
Lwls2D(
bw,
kern = "epan",
xin,
yin,
win = NULL,
xout1 = NULL,
xout2 = NULL,
xout = NULL,
subset = NULL,
crosscov = FALSE,
method = ifelse(kern == "gauss", "plain", "sort2")
)a p1 by p2 matrix of fitted values if xout is not specified. Otherwise a vector of length p corresponding to the rows of xout.
A scalar or a vector of length 2 specifying the bandwidth.
Kernel used: 'gauss', 'rect', 'gausvar', 'epan' (default), 'quar'.
An n by 2 data frame or matrix of x-coordinate.
A vector of y-coordinate.
A vector of weights on the observations.
a p1-vector of first output coordinate grid. Defaults to the input gridpoints if left unspecified.
a p2-vector of second output coordinate grid. Defaults to the input gridpoints if left unspecified.
alternative to xout1 and xout2. A matrix of p by 2 specifying the output points (may be inefficient if the size of xout is small).
a vector with the indices of x-/y-/w-in to be used (Default: NULL)
using function for cross-covariance estimation (Default: FALSE). FALSE for auto-covariance estimation and
TRUE for two-dimensional local linear kernel smoothing or cross-covariance estimation.
For auto-covariance estimation (i.e., when crosscov is FALSE), xout1 and xout2 should be the same.
should one try to sort the values xin and yin before using the lwls smoother? if yes ('sort2' - default for non-Gaussian kernels), if no ('plain' - fully stable; de)