Skip to contents

Computes the tricube weight function on a vector of distance values (d), based on a half-window width of h, and returns a vector of weights that range from zero to 1.

Usage

triCube(d, h)

Arguments

d

numeric vector of distances from the point of estimation to the given sample value

h

numeric value, the half-window width, measured in the same units as d

Value

w numeric vector of weights, all 0<=w<=1

Details

See Cleveland, W. S. (1979). Robust locally weighted regression and smoothing scatterplots, JASA, 74, 829-836

Examples

 h<-10
 d<-c(-11,-10,-5,-1,-0.01,0,5,9.9,10,20)
 triCube(d,h)
#>  [1] 0.000000e+00 0.000000e+00 6.699219e-01 9.970030e-01 1.000000e+00
#>  [6] 1.000000e+00 6.699219e-01 2.620072e-05 0.000000e+00 0.000000e+00