ml_clust {lazy.cluster}R Documentation

Normal Mixture Clustering

Description

This function uses mxnormal function to perform Maximum Likelihood cluster analysis under multivariate normal model.
The main purpose of this function is to be used from "dhclust" function.

Usage

ml_clust(x, ncl = 2, maxiter = 50, eps = 1e-06, nstart = 0,
  method = 1, init_method = "ward.D2", print = 0)

Arguments

x

obs x variable data frame

ncl

# of clusters to be sought

maxiter

max # of iterations

eps

criterion for convergence of ML

nstart

= 0 to use hclust solution as the initial
= n to use n random initail by kmeans to find the best solution

method

= 1 to use mxnormal, = 2 to use EMCluster::emcluster
(method = 2 deleted.)

init_method

method for clust

print

= 1 or 2 to print some result

Details

This version returns the ML cluster result only if
nrow(x) > 2*ncl
where 2 stands for the minimum obs to calculate variance.
Otherwise, LS solution will be sought by stat::kmeans function.

Value

A class "emret" object to which
betweenss, totss, tot.withinss
are added.

Examples

resml <- ml_clust( circle2, 5, print=2 )


[Package lazy.cluster version 0.1.2 Index]