Multiple.Random.fuzzy.kmeans: Multiple Random fuzzy-k-means clustering

In clusterv: Assessment of Cluster Stability by Randomized Maps

Multiple Random fuzzy-k-means clustering

Description

Multiple Random fuzzy-k-means clusterings are computed using random projections of data. The crisp clustering is obtained by defuzzyfication via the nearest crisp clustering: each example is assigned to the cluster for which it has the largest membership. The base fuzzy algorithm used is fanny of the cluster package. It assumes that the label of the examples are integers starting from 1 to ncol(M). Several randomized maps may be used: RS, PMO, Normal and Achlioptas random projections

Usage

Multiple.Random.fuzzy.kmeans(M, dim, pmethod = "PMO", c = 3, n = 50, 
                             scale = TRUE, seed = -1, distance = "euclidean")

Arguments

M	matrix of data: rows are variables and columns are examples
dim	subspace dimension
pmethod	projection method. It must be one of the following: "RS" (random subspace projection) "PMO" (Plus Minus One random projection) "Norm" (normal random projection) "Achlioptas" (Achlioptas random projection)
c	number of clusters
n	number of RS projections
scale	if TRUE randomized projections are scaled (default)
seed	numerical seed for the random generator
distance	it must be one of the two: "euclidean" (default) or "pearson" (that is 1 - Pearson correlation)

Value

a list of the n clusterings. Each clustering is a list of vectors, and each vector represents a single cluster. The elements of the vectors are integers that corresponds to the number of the columns (examples) of the matrix M of the data.

Author(s)

Giorgio Valentini [e-mail]

Examples

# Multiple (20) fuzzy-k-means clusterings using Normal projections. 
M <- generate.sample0(n=10, m=2, sigma=1, dim=800)
l.norm <- Multiple.Random.fuzzy.kmeans (M, dim=100, pmethod="Norm", c=3, n=20)
# The same as above, using Random Subspace projections.
l.RS <-  Multiple.Random.fuzzy.kmeans (M, dim=100, pmethod="RS", c=3,  n=20)
# The same as above, using PMO projections, but with the number of clusters set to 5
l.RS.PMO <-  Multiple.Random.fuzzy.kmeans (M, dim=100, pmethod="PMO", c=5, n=20)

clusterv documentation built on June 8, 2025, 10:21 a.m.