| equivKernel {locpol} | R Documentation |
Computes the Equivalent kernel for the local polynomial estimation.
equivKernel(kernel,nu,deg,lower=dom(kernel)[[1]],upper=dom(kernel)[[2]],
subdivisions=25)
nu |
Orders of derivative to estimate. |
deg |
Degree of Local polynomial estimator. |
kernel |
Kernel used to perform the estimation, see Kernels |
lower, upper |
Itegration limits. |
subdivisions |
the maximum number of subintervals. |
The definition of the Equivalent kernel for the local polynomial
estimation can be found in page 64 in Fan and Gijbels(1996). The
implementation uses computeMu to compute matrix S and then
returns a function object
Returns a vector whose components are the equivalent kenrel used to
compute the local polynomial estimator for the derivatives in nu.
Jorge Luis Ojeda Cabrera.
Fan, J. and Gijbels, I. Local polynomial modelling and its applications/. Chapman & Hall, London (1996).
## Some kernels and equiv. for higher order
## compare with p=1
curve(EpaK(x),-3,3,ylim=c(-.5,1))
f <- equivKernel(EpaK,0,3)
curve(f(x),-3,3,add=TRUE,col="blue")
curve(gaussK(x),-3,3,add=TRUE)
f <- equivKernel(gaussK,0,3)
curve(f(x),-3,3,add=TRUE,col="blue")
## Draw several Equivalent locl polynomial kernels
curve(EpaK(x),-3,3,ylim=c(-.5,1))
for(p in 1:5){
curve(equivKernel(gaussK,0,p)(x),-3,3,add=TRUE)
}