stepCwr: Function to estimate CWR models via multiple EM algorithm...

In cwm: Cluster Weighted Models by EM algorithm

Description

This function iteratens nIter times a single estimation of CWR models by cwrEm function. Then the one that has best logLikelihood is chosen.

Usage

stepCwr(X, Y, nc, prop = 0.1, nIter = 10, changeTrainingSet = FALSE)

Arguments

X	X data vector
Y	Y data vector
nc	number of clusters.
prop	Proportion of samples. Default 0.1.
nIter	Number of iteration. Default 10.
changeTrainingSet	Boolean. If TRUE the training set is changed.

Details

This function allows the estimation of models where the structure of the data set lies to probable convergence problems.

Value

A cwr object.

Note

Uses try.

Author(s)

Giorgio Spedicato

References

Murphy, Bettina.

See Also

cwrEm

Examples

data(geyser)
x=geyser[,1]
y=geyser[,2]                            
ciao=stepCwr(x,y,nc=2)

cwm documentation built on May 1, 2019, 9:19 p.m.