Kernel functional estimate for 1- to 6-dimensional data.
kfe(x, G, deriv.order, inc=1, binned=FALSE, bin.par, bgridsize, deriv.vec=TRUE,
add.index=TRUE, verbose=FALSE)
Hpi.kfe(x, nstage=2, pilot, pre="sphere", Hstart, binned=FALSE,
bgridsize, amise=FALSE, deriv.order=0, verbose=FALSE, optim.fun="nlm")
Hpi.diag.kfe(x, nstage=2, pilot, pre="scale", Hstart, binned=FALSE,
bgridsize, amise=FALSE, deriv.order=0, verbose=FALSE, optim.fun="nlm")
hpi.kfe(x, nstage=2, binned=FALSE, bgridsize, amise=FALSE, deriv.order=0)vector/matrix of data values
number of stages in the plug-in bandwidth selector (1 or 2)
"dscalar" = single pilot bandwidth (default) "dunconstr" = single unconstrained pilot bandwidth
"scale" = pre.scale, "sphere" = pre.sphere
initial bandwidth matrix, used in numerical optimisation
flag for binned estimation. Default is FALSE.
vector of binning grid sizes
flag to return the minimal scaled PI value
derivative order
flag to print out progress information. Default is FALSE.
optimiser function: one of nlm or optim
pilot bandwidth matrix
0=exclude diagonal, 1=include diagonal terms in kfe calculation
binning parameters - output from binning
flag to compute duplicated partial derivatives in the vectorised form. Default is FALSE.
flag to ouput derivative indices matrix. Default is true.
Plug-in bandwidth matrix for \(r\)-th order kernel functional estimator.
Hpi.kfe is the optimal plug-in bandwidth for \(r\)-th order kernel functional estimator
based on the unconstrained pilot selectors of Chacon & Duong (2010).
hpi.kfe is the 1-d equivalent, using the formulas from
Wand & Jones (1995, p.70).
kfe does not usually need to be called explicitly by the user.
Chacon, J.E. & Duong, T. (2010) Multivariate plug-in bandwidth selection with unconstrained pilot matrices. Test. 19, 375-398.
Wand, M.P. & Jones, M.C. (1995) Kernel Smoothing. Chapman & Hall/CRC, London.