Displaying 1 result from an estimated 1 matches for "seleciton".
2007 May 11
1
model seleciton by leave-one-out cross-validation
Hi, all
When I am using mle.cv(wle), I find a interesting problem: I can't do
leave-one-out cross-validation with mle.cv(wle). I will illustrate the
problem as following:
> xx=matrix(rnorm(20*3),ncol=3)
> bb=c(1,2,0)
> yy=xx%*%bb+rnorm(20,0,0.001)+0
> summary(mle.cv(yy~xx,split=nrow(xx)-1,monte.carlo=2*nrow(xx),verbose=T),
num.max=1)[[1]]
mle.cv: dimension of the split subsample