clustersim: clustersim: A cluster simulator for testing clustering...

In M3C: Monte Carlo Reference-based Consensus Clustering

Description

clustersim: A cluster simulator for testing clustering algorithms

Usage

clustersim(n, n2, r, K, alpha, wobble, redp = NULL, print = FALSE,
  seed = NULL)

Arguments

n	Numerical value: The number of samples, it must be square rootable
n2	Numerical value: The number of features
r	Numerical value: The radius to define the initial circle (use approx n/100)
K	Numerical value: How many clusters to simulate
alpha	Numerical value: How far to pull apart the clusters
wobble	Numerical value: The degree of noise to add to the sample co ordinates
redp	Numerical value: The fraction of samples to remove from one cluster
print	Logical flag: whether to print the PCA into current directory
seed	Numerical value: fixes the seed if you want to repeat results

Value

A list: containing 1) matrix with simulated data in it

Examples

res <- clustersim(225, 900, 8, 4, 0.75, 0.025, redp = NULL, seed=123)

Example output

***clustersim***
computing simulated data using linear combination...
calculating PCs of data and performing kmeans...
computing centroids of clusters and pulling apart...
transforming pulled apart PC co ordinates back to dataset...
finished.

M3C documentation built on Nov. 8, 2020, 6:50 p.m.