| Library: | cafpe |
| See also: | tp/cafpe/fvllc |
| Quantlet: | dencp | |
| Description: |
Quantlet for multivariate density estimation using
kernel estimation using C++ routines via a DLL.
|
| Usage: | dh = dencp(x,h,xest,loo,kernden) | |
| Input: | ||
| ATTENTION: | this quantlet requires to open locling.dll, density.dll (NT)
or locling.so, denc.so (UNIX). This can be done with the quantlet cafpeload or
directly with
garb = dlopen (" | |
| x | (n x d) matrix of observations | |
| h | (d x 1) vector of bandwidths | |
| xest | (nest x d) matrix of points at which the density has to be computed | |
| loo | scalar, parameter to control leave-one-out: 0 = no leave-one-out, 1 = leave-one-out | |
| kernden | character variable for selecting kernel function: "gaussian": Gaussian kernel | |
| Output: | ||
| dh | (nxest x 1) estimated density at xest | |
pathcafpe = "tp/cafpe/"
library("xplore")
func(pathcafpe + "cafpeload.xpl")
cafpeload(pathcafpe)
n = 100
x = normal(n,1)
loo = 0
kerntype = "gaussian"
h = hsilv(sqrt(var(x)),1,n,kerntype)
xs = sort(x)
dh = dencp(x,h,xs,loo,kerntype)
plotoneline(xs~dh, "density estimate", " ", " ")
| Library: | cafpe |
| See also: | tp/cafpe/fvllc |