Back Projected Kernel

Share:

Description

Evaluate a d dimensional back projected kernel.

Usage

1
bpk(X, bandwidths, a = 1.0, kernel = dnorm, ...)

Arguments

X

a numeric matrix of dimension N by d who's rows contain the points where the kernel will be evaluated.

bandwidths

a list with elements alphas, the set of projection directions, and lambdas, their associated bandwidth estimates.

a

a single numeric value containing a common scaling parameter.

kernel

a function for evaluating the univariate kernel.

...

additional arguments are ignored.

Value

a numeric vector where element i contains the value of the kernel evaluated at row i of X.

Author(s)

Kjell Konis kjell.konis@me.com