removeDuplicateFeatures {Mercator} | R Documentation |
The removeDuplicateFeatures
function removes duplicate columns from
a binaryMatrix
object in the Mercator
package.
removeDuplicateFeatures(object)
object |
An object of class |
In some analyses, it may be desirable to remove duplicate features to collapse a group of identical, related events to a single feature, to prevent overweighting when clustering.
Removal of duplicate features is not required for performance of the
binaryMatrix
or Mercator
objects and associated functions.
The history
slot of the binaryMatrix
object documents removal of
duplicate features.
Future versions of this package may include functionality to store the identities of duplicate features removed.
Returns an object of class binaryMatrix
with duplicate columns removed.
Transposing the binaryMatrix
can allow the removeDuplicateFeatures
function to be applied to both features and observations, if desired.
Features containing exclusively 0s or 1s may interfere with performance of
removeDuplicateFeatures
.
Kevin R. Coombes <krc@silicovore.com
my.matrix <- matrix(rbinom(50*100, 1, 0.15), ncol=50) my.matrix <- cbind(my.matrix, my.matrix[, 1:5]) # add duplicates dimnames(my.matrix) <- list(paste("R", 1:100, sep=''), paste("C", 1:55, sep='')) my.binmat <- BinaryMatrix(my.matrix) dim(my.binmat) my.binmat <- removeDuplicateFeatures(my.binmat) dim(my.binmat)