Graphical Markov Models with Mixed Graphs
Description
Provides functions for defining
mixed graphs containing three types of edges, directed,
undirected and bi-directed, with possibly multiple edges.
These graphs are useful because they capture fundamental
independence structures in multivariate distributions
and in the induced distributions after marginalization
and conditioning.
The package is especially concerned with Gaussian graphical
models for
(i) ML estimation for directed acyclic graphs, undirected and
bi-directed graphs and ancestral graph models
(ii) testing several conditional independencies
(iii) checking global identification of DAG Gaussian models
with one latent variable
(iv) testing Markov equivalences and generating Markov
equivalent graphs of specific types.