Tools

• Name:BayesBuilder (3)
  
• URL: http://www.snn.ru.nl/nijmegen/index.php?option=com_content&view=article&id=89&Itemid=212
• Handles directed, undirected, or chain (mixed) graphs:  I think it support directed.
  
• Learns parameters: No
• Learns structure: No
• Has a GUI: Yes.
• Platform (Windows, Linux, Mac): Available for Windows and Linux.
• Limitations/Concerns: Doesn't learn parameters or structures. BayesBuilder is available for non-commercial purposes.
  
• 
  
• Name: Bayesialab (3)
  
• URL: http://www.bayesia.com/en/products/bayesialab
• Handles directed, undirected, or chain graphs:  it supports only Directed.
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI: Yes
• Platform (Windows, Linux, Mac): As it is based on java, so I think it can run on any os installed with JRE
• Limitations/Concerns:  30 days trial
• 
  
• Name: Bayesian Knowledge Discoverer / Bayesware Discoverer (2)
  
• URL: http://kmi.open.ac.uk/projects/bkd/
• Handles directed, undirected, or chain (mixed) graphs: directed
• Learns parameters: yes
• Learns structure: yes
• Has a GUI: yes
• Platform (Windows, Linux, Mac): Windows, Unix and Macintosh
• Limitations/Concerns: Bayesian Knowledge Discoverer  is a free software. Bayesware Discoverer is a commercial version. A 30-days trial version is free of charge.
  
• 
  
• Name: Bayesian Network tools in Java (BNJ)
  
• URL: http://bnj.sourceforge.net/
• Handles directed, undirected, or chain graphs: only directed
• Learns parameters: No
• Learns structure: Yes
• Has a GUI: Yes
• Platform (Windows, Linux, Mac): JVM
• Limitations/Concerns: Free
• 
  
• Name: Bayesware (2)
  
• URL: http://bayesware.com/
  
• Handles directed, undirected, or chain (mixed) graphs: Directed only i think
  
• Learns parameters: Yes
  
• Learns structure: Yes
  
• Has a GUI: Yes
  
• Platform (Windows, Linux, Mac): Windows
  
• Limitations/Concerns: There is a free student edition for 30 days use.  After that you have to buy the full version.  "Cannot handle databases larger than 700 cases (records)." -- student version
• 
  
• Name: Bayes Net Toolbox for Matlab
  
• URL: http://code.google.com/p/bnt/
  
• Handles directed, undirected, or chain graphs: I think it supports only Directed.
• Learns parameters: Yes
  
• Learns structure: Yes
  
• Has a GUI: Yes
  
• Platform (Windows, Linux, Mac): Matlab
  
• Limitations/Concerns: None that I know.
• 
  
• 
  
• Name:B-Course
  
• URL: http://b-course.cs.helsinki.fi/obc/
• Handles directed, undirected, or chain (mixed) graphs: Directed
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI: Yes, runs on a browser.
• Platform (Windows, Linux, Mac): Runs on their server accessed using a web browser
• Limitations/Concerns: Size of the model is limited.
• 
  
• Name: Belief Network Constructor
  
• URL: http://webdocs.cs.ualberta.ca/~jcheng/bnsoft.htm
  
• Handles directed, undirected, or chain (mixed) graphs: Directed
  
• Learns parameters: Yes
  
• Learns structure: Yes, I think so
  
• Has a GUI: Yes
  
• Platform (Windows, Linux, Mac): Windows
  
• Limitations/Concerns: Last updated in 2001.
• 
  
• Name: Belief Network Power Constructor
  
• URL: http://webdocs.cs.ualberta.ca/~jcheng/bnpc.htm
• Handles directed, undirected, or chain (mixed) graphs: I think it support directed.
• Learns parameters: possible
• Learns structure: Yes
• Has a GUI:  Yes
• Platform (Windows, Linux, Mac): Windows
• Limitations/Concerns: free software
• 
  
• Name:Belief and Decision Networks 5.1.9
  
• URL: http://aispace.org/bayes
• Handles directed, undirected, or chain (mixed) graphs: Directed
• Learns parameters: Yes
• Learns structure: No
• Has a GUI: Yes.
• Platform (Windows, Linux, Mac): Java applet
• Limitations/Concerns: Nil
• 
  
• Name: Elvira
  
• URL: http://leo.ugr.es/elvira/
• Handles directed, undirected, or chain graphs: only directed
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI: Yes
• Platform (Windows, Linux, Mac): JVM
• Limitations/Concerns: Free
• 
  
• Name: FastInf
  
• URL: http://compbio.cs.huji.ac.il/FastInf/fastInf/FastInf_Homepage.html
• Handles directed, undirected, or chain graphs: I think it supports only Undirected.
• Learns parameters: Yes
• Learns structure: Yes, for template Markov Network
• Has a GUI: No
• Platform (Windows, Linux, Mac): Linux and Mac
• Limitations/Concerns: No Windows support
• 
  
• Name: "Flexible Bayesian Modeling and Markov Chain Sampling"
  
• URL: http://www.cs.toronto.edu/~radford/fbm.software.html
• Handles directed, undirected, or chain (mixed) graphs: appears to be directed only
• Learns parameters: Yes
• Learns structure: can't tell
• Has a GUI: No
• Platform (Windows, Linux, Mac): UNIX
• Limitations/Concerns: Open Source
• 
  
• Name: GeNIe and SMILE (its 2 tools, one for the graphical interface and the other for Modeling, Inference and Learning) (2)
  
• URL: http://genie.sis.pitt.edu/about.html
• Handles directed, undirected, or chain graphs: Only Directed
• Learns parameters: Yes / No
  
• Learns structure: Yes / No
  
• Has a GUI: Yes
• Platform: Windows, Linux and Mac. And it is free (requires registration).
• Limitations/Concerns: None that I know.
• 
  
• Name: GMTK
  
• URL: http://ssli.ee.washington.edu/~bilmes/gmtk/
• Handles directed, undirected, or chain (mixed) graphs: directed
• Learns parameters: yes
• Learns structure: no
• Has a GUI: no
• Platform (Windows, Linux, Mac): Linux/Solaris/Cygwin
• Limitations/Concerns: closed source, freely avaiable
• 
  
• Name: Hugin
  
• URL: http://www.hugin.com/
  
• Handles directed, undirected, or chain graphs: I think it supports only Directed.
  
• Learns parameters: Yes
  
• Learns structure: Yes
  
• Has a GUI: Yes
  
• Platform (Windows, Linux, Mac): Windows, Linux, Solaris, Mac.
  
• Limitations/Concerns: The full version is not free. The demo version is limited to 50 variables and 500 cases.
• 
  
• 
  
• 
  
• Name: JAGS (Just Another Gibbs Sampler)
  
• URL: http://www-fis.iarc.fr/~martyn/software/jags/
• Handles directed, undirected, or chain (mixed) graphs: directed
• Learns parameters: yes
• Learns structure: no
• Has a GUI: no
• Platform (Windows, Linux, Mac): UNIX
• Limitations/Concerns: port/clone of OpenBUGS, open source, ships in Debian/Ubuntu
• 
  
• Name: JBNC Toolbox-Weka
  
• URL: http://www.cs.waikato.ac.nz/ml/weka/ http://jbnc.sourceforge.net/
• Handles directed, undirected, or chain graphs: Bayesian Network only.
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI: Yes
• Platform (Windows, Linux, Mac): With support of JVM 1.6
• Limitations/Concerns: Weka provides a open source platform for data mining. With JBNC Toolkit, we can have different implemented algorithmsto training, testing, and applying BN classifiers.
• 
  
• Name: JProGraM - PRObabilistic GRAphical Model in Java
  
• URL: http://www.dii.unisi.it/~freno/JProGraM.html
• Handles directed, undirected, or chain (mixed) graphs: I think it support directed.
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI:  No
• Platform (Windows, Linux, Mac): Tested on several Linux distributions
• Limitations/Concerns: Released under the GNU
• 
  
• Name: MALLET + GRMM package (Java)
  
• URL: http://mallet.cs.umass.edu/ and  http://mallet.cs.umass.edu/grmm/index.php
• Handles directed, undirected, or chain (mixed) graphs: directed and undirected; not sure about mixed. Known mainly for CRF
• Learns parameters: Yes
• Learns structure: I think yes
• Has a GUI: No
• Platform (Windows, Linux, Mac): Written in Java so should be good for all platforms
• Limitations/Concerns: "GRMM can handle continuous features, but only discrete variables."
• 
  
• Name: MensXmachina
  
• URL: http://www.mensxmachina.org/software/pgm-toolbox/
• Handles directed, undirected, or chain (mixed) graphs: I think it support directed.
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI:  Yes
• Platform (Windows, Linux, Mac): Windows
• Limitations/Concerns: toolbox of matlab
• 
  
• Name: MIM
  
• URL: http://www.hypergraph.dk/
• Handles directed, undirected, or chain graphs: all
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI: Yes
• Platform (Windows, Linux, Mac): Windows platform only
• Limitations/Concerns: Not free; Up to 52 variables
• 
  
• Name: MSBNX
  
• URL: http://research.microsoft.com/en-us/um/redmond/groups/adapt/msbnx/
• Handles directed, undirected, or chain graphs:  it supports only Directed.
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI: Yes
• Platform (Windows, Linux, Mac): Windows 98, Windows 2000, and Windows XP.
• Limitations/Concerns:  None that I know
• 
  
• Name: Netica (3)
  
• URL: http://www.norsys.com/netica.html
• Handles directed, undirected, or chain graphs: only directed.
• Learns parameters: Yes
• Learns structure: No / Yes
  
• Has a GUI: Yes
• Platform: Windows, Mac.
• Limitations/Concerns: The full version is not free (285$ for students). The demo version is limited to model size (without specifiying). It can learn from a dataset that have missing values.
  
• 
  
• Name: Octave
  
• URL: http://www.gnu.org/software/octave/
• Handles directed, undirected, or chain graphs: It can run matlab code, and has some package itself as well.
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI: No
• Platform (Windows, Linux, Mac): Linux
• Limitations/Concerns: None that I know.
• 
  
• Name: OpenBUGS
  
• URL: http://www.openbugs.info
• Handles directed, undirected, or chain (mixed) graphs: directed
• Learns parameters: yes
• Learns structure: no
• Has a GUI: no
• Platform (Windows, Linux, Mac): Windows
• Limitations/Concerns: open source
• 
  
• Name: PMTK3
  
• URL: http://code.google.com/p/pmtk3/
• Handles directed, undirected, or chain graphs: All
• Learns parameters: Yes
• Learns structure: With MATLAB
• Has a GUI: Yes
• Platform (Windows, Linux, Mac): With support of MATLAB
• Limitations/Concerns: It should be run with MATLAB.
• 
  
• Name: Pulcinella (2)
  
• URL: http://iridia.ulb.ac.be/pulcinella/Welcome.html
• Handles directed, undirected, or chain (mixed) graphs: based on the framework of valuation systems
• Learns parameters: no
• Learns structure: no
• Has a GUI: no
• Platform (Windows, Linux, Mac): Solaris (Sun Unix) and Macintosh
• Limitations/Concerns: requires a Lisp system.
• 
  
• Name: Tetrad
  
• URL: http://www.phil.cmu.edu/projects/tetrad/index.html
• Handles directed, undirected, or chain (mixed) graphs: I think it support directed.
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI:  possible
• Platform (Windows, Linux, Mac): java based
• Limitations/Concerns: free beta version
• 
  
• Name: UGM
  
• URL: http://www.cs.ubc.ca/~schmidtm/Software/UGM.html
• Handles directed, undirected, or chain graphs: I think it supports only Undirected.
• Learns parameters: Yes
• Learns structure: No
• Has a GUI: I think no
• Platform (Windows, Linux, Mac): Matlab
• Limitations/Concerns: It just provides some Matlab functions.
• 
  
• Name: Weka
  
• URL: http://www.cs.waikato.ac.nz/ml/weka/
• Handles directed, undirected, or chain graphs: probably all.
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI: Yes
• Platform (Windows, Linux, Mac): Windows, Linux, Solaris, Mac.
• Limitations/Concerns: None that I know.
• 
  
• Name: WinMine (2)
  
• URL: http://research.microsoft.com/en-us/um/people/dmax/winmine/tooldoc.htm
• Handles directed, undirected, or chain (mixed) graphs: Both directed and undirected / directed.
  
• Learns parameters: Yes
• Learns structure: Yes
• Has a GUI: No, only command line operations. / Yes
  
• Platform (Windows, Linux, Mac): Windows
• Limitations/Concerns: Only works for Windows platforms and doesn't have a GUI.