40-generate.MixSim: Generate MixSim Examples for Testing

In pmclust: Parallel Model-Based Clustering using Expectation-Gathering-Maximization Algorithm for Finite Mixture Gaussian Model

Description

This function utilizes MixSim to generate sets of data for testing algorithms.

Usage

  generate.MixSim(N, p, K, MixSim.obj = NULL, MaxOmega = NULL,
                  BarOmega = NULL, PiLow = 1.0, sph = FALSE, hom = FALSE)

Arguments

N	total sample size across all S processors, i.e. sum over N.spmd is N.
p	dimension of data X.spmd, i.e. ncol(X.spmd).
K	number of clusters.
MixSim.obj	an object returned from MixSim.
MaxOmega	maximum overlap as in MixSim.
BarOmega	averaged overlap as in MixSim.
PiLow	lower bound of mixture proportion as in MixSim.
sph	sph as in MixSim.
hom	hom as in MixSim.

Details

If MixSim.obj is NULL, then BarOmega and MaxOmega will be used in MixSim to obtain a new MixSim.obj.

Value

A set of simulated data and information will be returned in a list variable including:

K	number of clusters, as the input
p	dimension of data X.spmd, as the input
N	total sample size, as the input
N.allspmds	a collection of sample sizes for all S processors, as the input
N.spmd	total sample size of given processor, as the input
X.spmd	generated data set with dimension with dimension N.spmd * p
CLASS.spmd	true id of each data, a vector of length N.spmd and has values from 1 to K
N.CLASS.spmd	true sample size of each clusters, a vector of length K
MixSim.obj	the true model where data X.spmd generated from

Author(s)

Wei-Chen Chen [e-mail] and George Ostrouchov.

References

Melnykov, V., Chen, W.-C. and Maitra, R. (2012) “MixSim: Simulating Data to Study Performance of Clustering Algorithms”, Journal of Statistical Software, (accepted).

Programming with Big Data in R Website: https://pbdr.org/

See Also

generate.basic.

Examples

## Not run: 
# Save code in a file "demo.r" and run in 4 processors by
# > mpiexec -np 4 Rscript demo.r

### Setup environment.
library(pmclust, quiet = TRUE)

### Generate an example data.
N <- 5000
p <- 2
K <- 2
data.spmd <- generate.MixSim(N, p, K, BarOmega = 0.01)
X.spmd <- data.spmd$X.spmd

### Run clustering.
PARAM.org <- set.global(K = K)          # Set global storages.
# PARAM.org <- initial.em(PARAM.org)    # One initial.
PARAM.org <- initial.RndEM(PARAM.org)   # Ten initials by default.
PARAM.new <- apecma.step(PARAM.org)     # Run APECMa.
em.update.class()                       # Get classification.

### Get results.
N.CLASS <- get.N.CLASS(K)
comm.cat("# of class:", N.CLASS, "\n")
comm.cat("# of class (true):", data.spmd$N.CLASS.spmd, "\n")

### Quit.
finalize()

## End(Not run)

pmclust documentation built on Feb. 11, 2021, 5:05 p.m.