[R-pkgs] depmixS4 version 1.3-0 on CRAN

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

[R-pkgs] depmixS4 version 1.3-0 on CRAN

Ingmar Visser
Package news (see below for general description of functionality)

depmixS4 version 1.3-0 has been released on CRAN. See the NEWS file
for an overview of all changes. The most important user-visible
changes are:

1) more compact pretty-printing of parameters in print/summary of
(dep)mix objects (following lm/glm style of presenting results)

2) some speed improvements in the EM algorithm, most notable in large
data/models

3) EM has an optional argument to use the classification likelihood
instead of the usual likelihood; this can be useful as a means of
starting value generation; use with caution as results are often
unstable.

Best, happy mixing, Ingmar & Maarten



Package general information

depmixS4 is a framework for specifying and fitting dependent mixture
models, otherwise known as hidden or latent Markov models.
Optimization is done with the EM algorithm or optionally with Rdonlp2
when (general linear (in-)equality) constraints on the parameters need
to be incorporated.  Models can be fitted on (multiple) sets of
observations.  The response densities for each state may be chosen
from the GLM family, or a multinomial.  User defined response
densities are easy to add; for the latter an example is given for the
ex-gauss distribution as well as the multivariate normal distribution.

Mixture or latent class (regression) models can also be fitted; these
are the limit case in which the length of observed time series is 1
for all cases.

_______________________________________________
R-packages mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-packages

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.