wire.init.fit: Get the initial values

In andrewthomasjones/EMMIXcontrasts: Contrasts in Mixed Effects for EMMIX Model with Random Effects

Description

The routinnes to fit mixture models to the data and to obtain initial partition or initial values.

Usage

wire.init.fit(dat, X, qe, n, m, g, nkmeans, nrandom = 0)

Arguments

dat	The dataset, an n by m numeric matrix, where n is number of observations and m the dimension of data.
X	The design matrix.
qe	The number of columns of design matrix W.
n	The number of observations.
m	The number of variables.
g	The number of components in the mixture model.
nkmeans	An integer to specify the number of KMEANS partitions to be used to find the best initial values.
nrandom	An integer to specify the number of random partitions to be used to find the best initial values; the default value is 0.

Details

These functions are called internally.

Value

A list containing

pro	A vector of mixing proportions pi
beta	A numeric matrix with each column corresponding to the mean.
sigma.e	The covaraince of error
cluster	A vector of final partition
loglik	The loglikelihood at convergence
lk	A vector of loglikelihood at each EM iteration

See Also

emmixwire

andrewthomasjones/EMMIXcontrasts documentation built on June 26, 2019, 5:46 a.m.