MineICA (version 1.12.0)

selectContrib: Select contributing features/genes

Description

This function selects elements whose absolute scaled values exceed a given threshold.

Usage

selectContrib(object, cutoff, level, ...)

Arguments

object

Either an IcaSet object, or a list of projection vectors, e.g the list of feature or gene projections on each component.

cutoff

The threshold according to which the elements will be selected. Must be either of length 1 and the same treshold is applied to all components, or of length equal to the number of components in order to use a specific threshold for each component.

level

The level of the selection: either "genes" to select contributing genes using SByGene(icaSet), or "features" to select contributing features using S(icaSet).

...

...

Value

A list of projections restricted to the elements that are higher than cutoff.

Details

Each vector is first scaled and then only elements with an absolute scaled value higher than cutoff are kept.

Examples

## Not run: 
# ## load an example of icaSet
# data(icaSetCarbayo)
# 
# ##### =========
# ####  When arg 'object' is an IcaSet object
# ##### =========
# 
# ## select contributing genes
# selectContrib(object=icaSetCarbayo, cutoff=3, level="genes")
# 
# ## select contributing features
# selectContrib(object=icaSetCarbayo, cutoff=3, level="features")
# 
# ##### =========
# #### When arg 'object' is a list
# ##### =========
#  c1 <- rnorm(100); names(c1) <- 100:199
#  c2 <- rnorm(100); names(c2) <- 1:99
#  selectContrib(object=list(c1,c2), cutoff= 0.5)
# 
# ## select contributing features
# contribFlist <- selectContrib(Slist(icaSetCarbayo), 3)
# 
# ## select contributing genes
# contribGlist <- selectContrib(SlistByGene(icaSetCarbayo), 3)
# 
# 
# ## End(Not run)