diceR (version 0.6.0)

consensus_combine: Combine algorithms

Description

Combines results for multiple objects from consensus_cluster() and outputs either the consensus matrices or consensus classes for all algorithms.

Usage

consensus_combine(..., element = c("matrix", "class"))

Arguments

...

any number of objects outputted from consensus_cluster()

element

either "matrix" or "class" to extract the consensus matrix or consensus class, respectively.

Value

consensus_combine returns either a list of all consensus matrices or a data frame showing all the consensus classes

Details

This function is useful for collecting summaries because the original results from consensus_cluster were combined to a single object. For example, setting element = "class" returns a matrix of consensus cluster assignments, which can be visualized as a consensus matrix heatmap.

Examples

# NOT RUN {
# Consensus clustering for multiple algorithms
suppressWarnings(RNGversion("3.5.0"))
set.seed(911)
x <- matrix(rnorm(500), ncol = 10)
CC1 <- consensus_cluster(x, nk = 3:4, reps = 10, algorithms = "ap",
progress = FALSE)
CC2 <- consensus_cluster(x, nk = 3:4, reps = 10, algorithms = "km",
progress = FALSE)

# Combine and return either matrices or classes
y1 <- consensus_combine(CC1, CC2, element = "matrix")
str(y1)
y2 <- consensus_combine(CC1, CC2, element = "class")
str(y2)
# }