Set of PARAM: A Set of Parameters in Model-Based Clustering.
Description
This set of parameters are used in initialization, EM iterations, and
final convergent results. All share the same structure in a list variable.
Arguments
format
A list variable contains several parameters for computing.
Details
The elements of PARAM or PARAM.org are
ll{N number of observations
p dimension of each observation,
total number of variables
K number of clusters
ETA mixing proportion
log.ETA log of mixing proportion
MU centers, dim =
$p \times K$
SIGMA dispersions, a list containing $K$
elements, each element is a matrix,
dim = $p \times p$
U Choleski of SIGMA, the same size of
SIGMA
U.check checks of each elements of U,
length $K$
logL log likelihood
min.N.CLASS minimum number of elements in a
cluster (restrictions)
}
The model parameters are ETA, MU, and SIGMA, while
log.ETA, U, U.check, and min.N.CLASS
are only used in computing.
References
High Performance Statistical Computing (HPSC) Website:
http://thirteen-01.stat.iastate.edu/snoweye/hpsc/
Programming with Big Data in R Website:
http://r-pbd.org/