widals (version 0.6.2)

create.rm.ndx.ls: Cross-Validation Indices

Description

Create a list of vectors of indices to remove for k-fold cross-validation

Usage

create.rm.ndx.ls(n, xincmnt = 10)

Value

An unnamed list of integer (>0) vectors.

Arguments

n

Number of sites. A scalar integer.

xincmnt

How many cv folds, i.e., k.

Details

The name of the object produced by this function is commonly rm.ndx in this documentation. See MSS.snow for a reminder that this object is passed out-of-scope when using MSS.snow.

In this package rm.ndx is used by Hals.fastcv.snow and widals.snow; however, creating this object as a list using this function is only necessary when using widals.snow with cv=2 (i.e., 'true' cross-validation).

Examples

n <- 100
xincmnt <- 7
rm.ndx <- create.rm.ndx.ls(n=n, xincmnt=xincmnt)
rm.ndx

######## if we want randomization of indices:
n <- 100
xincmnt <- 7
rm.ndx <- create.rm.ndx.ls(n=n, xincmnt=xincmnt)

rnd.ndx <- sample(I(1:n))
for(i in 1:length(rm.ndx)) { rm.ndx[[i]] <- rnd.ndx[rm.ndx[[i]]] }
rm.ndx

## The function is currently defined as
function (n, xincmnt = 10) 
{
    rm.ndx.ls <- list()
    for (i in 1:xincmnt) {
        xrm.ndxs <- seq(i, n + xincmnt, by = xincmnt)
        xrm.ndxs <- xrm.ndxs[xrm.ndxs <= n]
        rm.ndx.ls[[i]] <- xrm.ndxs
    }
    return(rm.ndx.ls)
}