simPATHy (version 0.6)

fitSgraph: Estimate covariance matrix of a graphical model

Description

Fit a Gaussian Graphical Model or a Gaussian Bayesian Network by maximum likelihood.

Usage

fitSgraph(graph, S)

Arguments

graph

A directed or undirected graph represented as a graphNEL object.

S

A sample covariance matrix

Value

A covariance matrix with the independence constraints entailed by the graph.

Details

If graph is undirected it uses the Iterative Proprotional Fitting algoritm (qpgraph package). If graph is directed it uses Iterative Conditional Fitting (ggm package).

References

Drton, M. & Richardson, T. S. (2003). A new algorithm for maximum likelihood estimation in Gaussian graphical models for marginal independence. Proceedings of the Ninetheen Conference on Uncertainty in Artificial Intelligence, 184-191.

Whittaker, J. Graphical models in applied multivariate statistics. Wiley, 1990.

See Also

icfmag, qpIPF