Daniel@0:

Visualizing graph structures in matlab

Daniel@0: Daniel@0: We discuss some methods for visualizing graphs/ networks, including automatic Daniel@0: layout of the nodes. Daniel@0: We assume the graph is represented as an adjacency matrix. Daniel@0: If using BNT, you can access the DAG using Daniel@0:

Daniel@0: G = bnet.dag;
Daniel@0:

Daniel@0: Daniel@0:

Matlab's biograph function

Daniel@0: Daniel@0: The Mathworks computational biology toolbox Daniel@0: has many useful graph related functions, including visualization. Daniel@0:
Daniel@0: Click Daniel@0: Daniel@0: here Daniel@0: for a demo. Daniel@0: Daniel@0: Daniel@0: Daniel@0:

Cemgil's draw_graph

Daniel@0: Daniel@0: You can visualize an arbitrary graph (such as one learned using the Daniel@0: structure learning routines) with Matlab code written by Daniel@0: Daniel@0: Ali Taylan Cemgil Daniel@0: from the University of Cambridge. Daniel@0: A modified version of this code Daniel@0: is here Daniel@0: (this is already bundled with BNT). Daniel@0: Just type Daniel@0:

Daniel@0: draw_graph(G);
Daniel@0:

Daniel@0: For example, this is the output produced on a Daniel@0: random QMR-like model: Daniel@0:

Daniel@0: Daniel@0:

Pajek

Daniel@0: Daniel@0: Pajek Daniel@0: is an excellent, free Windows program for graph layout. Daniel@0: Use adj2pajek2.m to convert a graph to the Daniel@0: Pajek file format. Daniel@0:
Daniel@0: Then Choose File->Network->Read from the menu. Daniel@0: Daniel@0:

AT&T Graphviz

Daniel@0: Daniel@0: graphhviz Daniel@0: is an Daniel@0: open-source graph visualization package from AT&T. Daniel@0: Use Daniel@0: graph_to_dot Daniel@0: to convert an adjacency matrix to Daniel@0: the AT&T file format (the "dot" format). Daniel@0: You then use dot to convert it to postscript: Daniel@0:

Daniel@0: graph_to_dot(G, 'filename', 'foo.dot');
Daniel@0: dot -Tps foo.dot -o foo.ps
Daniel@0: ghostview foo.ps &
Daniel@0: