get.segments: Subsampling segments
In BioMark: Find Biomarkers in Two-Class Discrimination Problems

Description Usage Arguments Value Author(s) See Also Examples

View source: R/segments.R

Provides combinations of samples to be left out in subsampling, with a maximum given by parameter max.seg.

1	get.segments(i1, i2 = NULL, oob.size = 1, max.seg = 100)

`i1`	either an index vector for objects in class 1, or a classification vector (factor, or numeric), from which the indices of both classes can be derived.
`i2`	if non-NULL, vector indexing objects in class 2.
`oob.size`	number of samples to be left out in every iteration. If one (the default), this corresponds to LOO subsampling.
`max.seg`	maximal number of segments to return. If null, all possible combinations are returned – this option is only possible if oob.size equals 1. If oob.size is larger, max.seg must be defined since the number of possibilities becomes too large for even very small numbers of objects.

Returns a matrix where the columns contain the numbers of the samples to be left out in the respective iterations.

Ron Wehrens

get.biom

i1 <- seq(1, 10, by = 2)
i2 <- seq(2, 15, by = 2)
get.segments(i1, i2)
get.segments(i1, i2, max.seg = 10)
get.segments(i1, i2, oob.size = 2, max.seg = 10)

I <- rep(1:2, c(5,6))
get.segments(I)
get.segments(I, max.seg = 15)