convergencePlot: Show EM Algorithm Convergence

In lcmix: Layered and chained mixture models

Description

Plot convergence of the log-likelihood to a maximum during execution of the EM algorithm for fitting mixture models. Optionally also show the Q-value, that is, the expected complete data log-likelihood.

Usage

convergencePlot(x, show.qval=FALSE, main="")

Arguments

x	an object of class mixmod or mdmixmod.
show.qval	logical; if TRUE, the Q-value will be plotted as well as the log-likelihood.
main	plot title.

Details

Log-likelihood is plotted vs. iteration in black. If show.qval is TRUE, Q-value is plotted in red.

Author(s)

Daniel Dvorkin

See Also

mixmod, mdmixmod, rocinfo; plot in package graphics.

Examples

## Not run: 
data(CiData)
fit <- mixmod(CiData$expression, 3)
convergencePlot(fit, TRUE, main="log-likelihood and Q-value")

## End(Not run)

lcmix documentation built on May 2, 2019, 6:49 p.m.