GMtoolbox

GMtoolbox : a Matlab toolbox for inference in Graphical Models

Quick Start

To quickly learn how to use the toolbox, see the QuickStart manual based on a Dynamic Bayesian Network (DBN) and a Coupled Hidden Markov Model (CHMM) problems.

GMtoolbox functions

For a detailed description, see the reference manual.

GM generation

gm_create_fgDefine a factor graph
gm_example_SprinklerGenerate a Bayesian Network toy example
gmdp_example_EcologyGenerate a toy example related to ecology
gmdp_example_DBNGenerate a Dynamic Bayesian Network
gmdp_example_CHMMGenerate a Coupled Hidden Markov Model
gm_create_fgGenerate a factor graph
gm_separate_fgSeparate connected sub factor graphs
 
GM inference

gm_infer_GBPGeneralized Belief Propagation (parent-to-child)
gm_rg_CVMGenerate a region graph with Cluster Variation Method (CVM)
gm_rg_BETHEGenerate a region graph with BETHE approximation
gm_rg_JTGenerate a region graph such that GBP corresponds to Junction Tree (JT)
gm_include_evidenceTake into account observations
gm_plot_belief_evolutionPlot the evolution of belief for a range of GBP iteration
 
Utilities

gm_check_fg Check the validity of a factor grah
gm_check_rg Check the validity of a region graph
gm_check_rg_JTCheck if a region graph is of the type given by gm_rg_JT function
gm_plot_fg Plot a factor grah
gm_plot_rg Plot a region grah
gm_read_fg Read a factor graph in UAI format
gm_read_evidence Read evidences in UAI format
gm_read_MAR Read marginals in UAI format
gm_read_BEL Read beliefs of regions in UAI format
gm_write_fg Write a factor graph in UAI format
gm_write_evidence Write evidences in UAI format
gm_write_MAR Write marginals in UAI format
gm_write_BEL Write beliefs of regions in UAI format