Library: | smoother |
See also: | Rdenbest denxbwcrit denxest |
Quantlet: | Rdenxest | |
Description: | evaluates a kernel estimate of an integrated squared density (derivative) using the normal kernel for a (vector of) bandwidth(s) h. |
Usage: | Rh = Rdenxest(der, x, h, diag) | |
Input: | ||
der | scalar, order of derivative der = 0,1,2,... | |
x | n x 1 vector of data points | |
h | p x 1 vector of bandwidths | |
diag | scalar, if set to 0, the diagonal terms are removed from the estimate, otherwise included | |
Output: | ||
Rh | p x 1 vector of the functional estimates |
library("smoother") library("xplore") randomize(0) n = 100 s = 2 diag = 1 h = #(0.1, 0.2, 0.3) {w,mu,sigma}=normalmixselect("Marron_Wand_3") x = normalmix(n,w,mu,sigma) Rh = Rdenxest(s,x,h,diag) h ~ Rh
kernel estimates for the integrated squared second density derivative for three different bandwidths are computed which are based on a sample of size 100 generated from a normal mixture example density: Contents of _tmp [1,] 0.1 1742 [2,] 0.2 111.71 [3,] 0.3 18.953