ISE_fstar {OSCV} | R Documentation |
fstar
.Computing the ISE function for the Gaussian density estimator obtained from a random sample of size n
generated from fstar
.
ISE_fstar(h, n)
h |
numerical vector of bandwidth values, |
n |
sample size (number of data points generated from |
The integrated squared error (ISE) is a measure of closeness of the Gaussian density estimate computed from a data set generated from fstar
to the true density.
The vector of values of the ISE function for the correponsing vector of h
values.
Savchuk, O.Y. (2017). One-sided cross-validation for nonsmooth density functions, arXiv:1703.05157.
## Not run:
dev.new()
harray=seq(0.05,1.5,len=1000)
ISEarray=ISE_fstar(harray,100)
h_ISE=round(harray[which.min(ISEarray)],digits=4)
dev.new()
plot(harray,ISEarray,lwd=3,'l',xlab="h",ylab="ISE",main="ISE(h)",cex.main=2,cex.lab=1.7,
cex.axis=1.7)
legend(0.35,ISEarray[5],legend=c("n=100",paste("h_ISE=",h_ISE)),cex=1.8,bty="n")
## End(Not run)