RMBCaux: RMBCaux

In RMBC: Robust Model Based Clustering

Description

Robust Model Base Clustering algorithm based on centers, a robust and efficient version of EM algorithm.

Usage

RMBCaux(
  Y,
  K,
  thetaOld.alpha,
  thetaOld.mu,
  thetaOld.sigma,
  max_iter,
  niterFixedPoint,
  tolerance,
  cutoff = 1 - 0.001
)

Arguments

Y	A matrix of size n x p.
K	The number of clusters.
thetaOld.alpha	The initial alpha: An array of K positive real numbers they must verify the condition sum(thetaOld.mu)== 1.
thetaOld.mu	The initial centers: A list with K elements, each of them is an array of length p.
thetaOld.sigma	The initial stcatter matrix: A list with K matrix, each of them has dimension p x p
max_iter	a maximum number of iterations used for the algorithm stopping rule
niterFixedPoint	the maximum number of iteration in the internal loop which computes sigma an mu separately. The default value is niterFixedPoint=1
tolerance	tolerance parameter used for the algorithm stopping rule
cutoff	optional argument for outliers detection - quantiles of chi-square to be used as a threshold for outliers detection, defaults to 0.999

Value

A list including the estimated K centers and labels for the observations

• centers: matrix of size K x p, with the estimated K centers.

• cluster: array of size n x 1 integers labels between 1 and K.

• tauPath: sequence of tau scale values at each iterations.

• Wni: numeric array of size n x 1 indicating the weights associated to each observation.

• emptyClusterFlag: a boolean value. True means that in some iteration there were clusters totally empty

• niter: number of iterations until convergence is achived or maximum number of iteration is reached

• didistance of each observation to its assigned cluster-center

RMBC documentation built on July 22, 2021, 9:07 a.m.