setEmOptions: Set Expectation and Maximization options.

In ClustMMDD: Variable Selection in Clustering by Mixture Models for Discrete Data

Description

Set Expectation and Maximization options.

Usage

setEmOptions(emOptions = list(epsi = NULL, typeSmallEM = NULL, typeEM = NULL,
  nberSmallEM = NULL, nberIterations = NULL, nberMaxIterations = NULL,
  putThreshold = NULL))

Arguments

emOptions

A list of options needed by the Expectation and maximization algorithm : epsi : In [1e-5, 1e-20], it is the upper bound of the relative increase in the log-likelihood. typeSmallEM : In c(0, 1, 2) : 0 = classic EM, 1 = SEM, 2 = CEM. typeEM : In c(0, 1, 2) : 0 = classic EM, 1 = SEM, 2 = CEM. nberSmallEM : The number of random parameter points from which to perform nberIterations EM runs. nberIterations : The number of iterations for each small EM. nberMaxIterations : The maximum number of iterations if EM algorithm converge hardly. putThreshold : If TRUE, the probabilities of levels are assumed to be positive in all clusters.

Details

Use setEmOptions() to set all options to default.

Author(s)

Wilson Toussile.

References

• Dominique Bontemps and Wilson Toussile (2013) : Clustering and variable selection for categorical multivariate data. Electronic Journal of Statistics, Volume 7, 2344-2371, ISSN.

• Wilson Toussile and Elisabeth Gassiat (2009) : Variable selection in model-based clustering using multilocus genotype data. Adv Data Anal Classif, Vol 3, number 2, 109-134.

See Also

EmOptions for getting the current EM options.

Examples

EmOptions()
setEmOptions(list(epsi = 1e-6))
EmOptions()
setEmOptions()	# To set default values
EmOptions()

ClustMMDD documentation built on May 2, 2019, 2:44 p.m.