unbbayes-users Mailing List for UnBBayes

Hello,
        Is there a way to calculate time consumed by a probabilistic query
in Unbbayes GUI ?



Thanks,
Vasim

hi all,
   I have downloaded the latest source codes from svn and build it.
But there is an error when testing.
testCostRatio(unbbayes.evaluation.WetGrassEvaluationTest)  Time
elapsed: 0.012 sec  <<< FAILURE!
java.lang.AssertionError: expected:<5.75E-4> but was:<17.433692932128906>
    at org.junit.Assert.fail(Assert.java:69)
    at org.junit.Assert.failNotEquals(Assert.java:314)
    at org.junit.Assert.assertEquals(Assert.java:163)
    at org.junit.Assert.assertEquals(Assert.java:178)
    at unbbayes.evaluation.WetGrassEvaluationTest.testCostRatio(WetGrassEvaluationTest.java:394)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:606)
    at org.junit.internal.runners.TestMethodRunner.executeMethodBody(TestMethodRunner.java:99)
    at org.junit.internal.runners.TestMethodRunner.runUnprotected(TestMethodRunner.java:81)
    at org.junit.internal.runners.BeforeAndAfterRunner.runProtected(BeforeAndAfterRunner.java:34)
    at org.junit.internal.runners.TestMethodRunner.runMethod(TestMethodRunner.java:75)
    at org.junit.internal.runners.TestMethodRunner.run(TestMethodRunner.java:45)
    at org.junit.internal.runners.TestClassMethodsRunner.invokeTestMethod(TestClassMethodsRunner.java:71)
    at org.junit.internal.runners.TestClassMethodsRunner.run(TestClassMethodsRunner.java:35)
    at org.junit.internal.runners.TestClassRunner$1.runUnprotected(TestClassRunner.java:42)
    at org.junit.internal.runners.BeforeAndAfterRunner.runProtected(BeforeAndAfterRunner.java:34)
    at org.junit.internal.runners.TestClassRunner.run(TestClassRunner.java:52)
    at org.apache.maven.surefire.junit4.JUnit4Provider.execute(JUnit4Provider.java:252)
    at org.apache.maven.surefire.junit4.JUnit4Provider.executeTestSet(JUnit4Provider.java:141)
    at org.apache.maven.surefire.junit4.JUnit4Provider.invoke(JUnit4Provider.java:112)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:606)
    at org.apache.maven.surefire.util.ReflectionUtils.invokeMethodWithArray(ReflectionUtils.java:189)
    at org.apache.maven.surefire.booter.ProviderFactory$ProviderProxy.invoke(ProviderFactory.java:165)
    at org.apache.maven.surefire.booter.ProviderFactory.invokeProvider(ProviderFactory.java:85)
    at org.apache.maven.surefire.booter.ForkedBooter.runSuitesInProcess(ForkedBooter.java:115)
    at org.apache.maven.surefire.booter.ForkedBooter.main(ForkedBooter.java:75)

Hi! Due to the amount of emails we have received recently, we have decided
to start using the forum available at SF. We created several forums, one for
each main topic: - Basic Features; - Probabilistic Networks; - Learning; -
Simulation; - Data Mining; - Metaphor; and - Papers. We ask that from now
on, if you have any comments or question you would like to send us, please
do so by posting at one of our forums
https://sourceforge.net/projects/unbbayes/forums. The main goal is to avoid
answering the same question more than once and to make sure users have the
feedback they need ASAP. This way they do not have to wait for us to answer
their question if it is already available in our forum. All the best,
UnBBayes Team

Hi!

The UnBBayes Team would like to kindly thank the Genie group (
http://genie.sis.pitt.edu/) for allowing us to use part of their tutorial in
our own. Most of the material explaining basic modeling, BN, etc was
borrowed from Genie's tutorial (
http://genie.sis.pitt.edu/wiki/GeNIe_Documentation).

Thanks again!!
UnBBayes Team

Hi! We are really excited with our newest addition to our web site! A lot of
people have complained about the lack of international (non portuguese)
documentation for UnBBayes and finally, with the help of professors and
students from George Mason University, we have published our official
tutorial in our wiki at
https://sourceforge.net/apps/mediawiki/unbbayes/index.php?title=Visit_the_Tutorial.
Please understand that this is the first version and we are counting with
your help to fix problems, make it clearer, and add explanations about other
features we have not covered in detail. I.e., let us know if you see a
problem or if you want to make a contribution. Besides the tutorial, we have
added presentations we have given before about UnBBayes in our web site at
http://unbbayes.sourceforge.net/presentations.html. Most presentations were
given by member of the Artificial Intelligence Group (GIA) from the
University of Brasilia (UnB). Finally, we have included some new videos in
our video tutorial section at
http://unbbayes.sourceforge.net/video_tutorial.html. We hope you enjoy!
After all, we did this, as we said before, due to the numerous requests we
received by e-mail and as an attempt to get even more people using UnBBayes
and developing plugins for the new plugin architecture we have developed.
All the best, UnBBayes Team

Hi all, It is our pleasure to announce the so promised version with plugin
support! Now the community will be able to contribute even more by creating
their own inference algorithm, resource translation, entire new module, and
much more by using our simple plugin framework. Download the UnBBayes Plugin
Framework Core today and try out all the different plugins already
available! Remember to download the Core first! The plugins are installed
just by unziping the .zip file on the plugins folder inside unbbayes core
folder. Also check out our new website at http://unbbayes.sourceforge.net.
There you can see the changes available on this new version. Besides that, a
lot of new video tutorials are available. We hope you enjoy the improvements
we have been working so hard on! All the best, UnBBayes Team

Hi all,

We are really proud to announce this new release of UnBBayes! This release
took a lot of people a great amount of time and effort.

As stated by ohloh.net website, it would cost 5,667,296.00 dollars (that is
right! more than 5 millions!!!) to hire a team to write this project
(UnBBayes) from scratch! And you have all that for free!!!

All we ask in return is that you be kind enough to write us a review at
https://sourceforge.net/projects/unbbayes/ and at
https://www.ohloh.net/p/unbbayes. We would also really appreciate if you
could register yourself as a UnBBayes user at
https://www.ohloh.net/p/unbbayes.

This version has a lot improvements, from algorithms to a new and great user
interface:
- New GUI (Young);
- Progress Bar (Aditya);
- Sampling (Rommel);
- Approximate Inference (Rommel);
- Dr. Laskey's SSBN Algorithm (Laecio, Shou, and Rommel);
- Dr. Chang's Classification Performance Evaluation (Rommel);
- Metaphor (lots of people from UnB/Brazil);
- UnBMiner (lots of people from UnB/Brazil);
- and more!

You can download it from our website at
https://sourceforge.net/projects/unbbayes/.

All the best,
UnBBayes Team

Hi all!
Recently we have received a lot of e-mails from users asking us to help them
make the best use of UnBBayes. Since UnBBayes has been developed by
Brazilians from the University of Brasilia (UnB), most of the dissertations,
thesis, and documentation we have are in Portuguese.

Unfortunately, most of UnBBayes' users do not speak Portuguese. Therefore,
we have been working with George Mason University (GMU) on
creating/translating these tutorials in/to English.

Part of this effort has been on creating video tutorial to help users get
familiar with UnBBayes even faster.

Now you can access these tutorials at our website
http://unbbayes.sourceforge.net/video_tutorial.html.

Cheers,
UnBBayes Team

Hi all!
Recently we have received a lot of e-mails from users asking us to help them
make the best use of UnBBayes. Since UnBBayes has been developed by
Brazilians from the University of Brasilia (UnB), most of the dissertations,
thesis, and documentation we have are in Portuguese.

Unfortunately, most of UnBBayes' users do not speak Portuguese. Therefore,
we have been working with George Mason University (GMU) on
creating/translating these tutorials in/to English.

Part of this effort has been on creating video tutorial to help users get
familiar with UnBBayes even faster.

Now you can access these tutorials at our website
http://unbbayes.sourceforge.net/video_tutorial.html.

Cheers,
UnBBayes Team

Hi Ivan,
I just saw this message today! I do not know why this message was not
forwarded to my e-mail. I am sorry for that!

I would guess that you downloaded our API, instead of downloading the
Framework. The API can only be used as a library, therefore, you cannot run
it. Besides that, to run a jar file you have to type:
>java -jar unbbayes2.jar

Nevertheless, I do not think it will work either, since it seems that this
jar file is just the API as I said before.

Try downloading one of our Framework versions and running it by double
clicking on the jar file or running the command above from the cmd.

Cheers,
Rommel Carvalho
*[Unbbayes-users] Error trying to launch unbbayes*
Exclude<https://sourceforge.net/mailarchive/exclude_list.php?action=addexclude&group_id=47519&forum_name=unbbayes-users&msg...@ma...>
From:
Ivan Sopov <moradan228@gm...> - 2009-08-04 19:40
Hello, here is the output of the launching from the cmd and
information about the JVM (it is the latest Sun JDK).
C:\>java unbbayes2.jar
Exception in thread "main" java.lang.NoClassDefFoundError: unbbayes2/jar
Caused by: java.lang.ClassNotFoundException: unbbayes2.jar
at java.net.URLClassLoader$1.run(Unknown Source)
at java.security.AccessController.doPrivileged(Native Method)
at java.net.URLClassLoader.findClass(Unknown Source)
at java.lang.ClassLoader.loadClass(Unknown Source)
at sun.misc.Launcher$AppClassLoader.loadClass(Unknown Source)
at java.lang.ClassLoader.loadClass(Unknown Source)
at java.lang.ClassLoader.loadClassInternal(Unknown Source)
Could not find the main class: unbbayes2.jar. Program will exit.

C:\>java -version
java version "1.6.0_14"
Java(TM) SE Runtime Environment (build 1.6.0_14-b08)
Java HotSpot(TM) Client VM (build 14.0-b16, mixed mode, sharing)

C:\>java -jar unbbayes2.jar
Exception in thread "main" java.lang.NoClassDefFoundError:
unbbayes/controller/M
ainController
at unbbayes.Main.main(Main.java:38)
Caused by: java.lang.ClassNotFoundException:
unbbayes.controller.MainController
at java.net.URLClassLoader$1.run(Unknown Source)
at java.security.AccessController.doPrivileged(Native Method)
at java.net.URLClassLoader.findClass(Unknown Source)
at java.lang.ClassLoader.loadClass(Unknown Source)
at sun.misc.Launcher$AppClassLoader.loadClass(Unknown Source)
at java.lang.ClassLoader.loadClass(Unknown Source)
at java.lang.ClassLoader.loadClassInternal(Unknown Source)
... 1 more


I've noticed that I've got a corrupted download from the sourceforge
as it is usual with firefox and reload it several times - at last I
got the correct jar file which I can extract, but error remains the
same. I've tried to javac that MainController.java but it seems rather
tricky for me from the cmd.
Hope to be useful, Ivan Sopov.

Hi all!
We have been working on creating some tutorials in English for UnBBayes. We
have received a lot of requests recently for that.

We are just finishing editing the tutorials to publish them in our wiki
page. Besides that, we have started creating a few video tutorials to help
the user get familiar with UnBBayes. I have just published the first one at
http://www.youtube.com/watch?v=ZO1y-q20GPo. This is just a brief overview of
UnBBayes. We will be publishing more videos in next few days/weeks.

Please, do not hesitate to send us questions you might have.

All the best,
UnBBayes Team

Hello, here is the output of the launching from the cmd and
information about the JVM (it is the latest Sun JDK).
C:\>java unbbayes2.jar
Exception in thread "main" java.lang.NoClassDefFoundError: unbbayes2/jar
Caused by: java.lang.ClassNotFoundException: unbbayes2.jar
        at java.net.URLClassLoader$1.run(Unknown Source)
        at java.security.AccessController.doPrivileged(Native Method)
        at java.net.URLClassLoader.findClass(Unknown Source)
        at java.lang.ClassLoader.loadClass(Unknown Source)
        at sun.misc.Launcher$AppClassLoader.loadClass(Unknown Source)
        at java.lang.ClassLoader.loadClass(Unknown Source)
        at java.lang.ClassLoader.loadClassInternal(Unknown Source)
Could not find the main class: unbbayes2.jar.  Program will exit.

C:\>java -version
java version "1.6.0_14"
Java(TM) SE Runtime Environment (build 1.6.0_14-b08)
Java HotSpot(TM) Client VM (build 14.0-b16, mixed mode, sharing)

C:\>java -jar unbbayes2.jar
Exception in thread "main" java.lang.NoClassDefFoundError: unbbayes/controller/M
ainController
        at unbbayes.Main.main(Main.java:38)
Caused by: java.lang.ClassNotFoundException: unbbayes.controller.MainController
        at java.net.URLClassLoader$1.run(Unknown Source)
        at java.security.AccessController.doPrivileged(Native Method)
        at java.net.URLClassLoader.findClass(Unknown Source)
        at java.lang.ClassLoader.loadClass(Unknown Source)
        at sun.misc.Launcher$AppClassLoader.loadClass(Unknown Source)
        at java.lang.ClassLoader.loadClass(Unknown Source)
        at java.lang.ClassLoader.loadClassInternal(Unknown Source)
        ... 1 more


I've noticed that I've got a corrupted download from the sourceforge
as it is usual with firefox and reload it several times - at last I
got the correct jar file which I can extract, but error remains the
same. I've tried to javac that MainController.java but it seems rather
tricky for me from the cmd.
Hope to be useful, Ivan Sopov.

I figured out what seems like what should be an unnecessary workaround for the problem that I've just posted.  By changing my code from

        for(int i = 0; i < NUM_RUNS; i++){
            network.initialize();

            if(i > 0) d.addLikeliHood(beliefs[i-1]);

            a.addFinding(aFindings[i]);
            b.addFinding(bFindings[i]);
            network.updateEvidences();
            float[] tempBeliefs = new float[c.getStatesSize()];
            for(int j = 0; j < c.getStatesSize(); j++){
                tempBeliefs[j] = c.getMarginalAt(j);
            }
            beliefs[i] = tempBeliefs;
        }

To 

        for(int i = 0; i < NUM_RUNS; i++){
            network.initialize();

            if(i > 0){
			d.addFinding(0);
			d.addLikeliHood(beliefs[i-1]);
		}

            a.addFinding(aFindings[i]);
            b.addFinding(bFindings[i]);
            network.updateEvidences();
            float[] tempBeliefs = new float[c.getStatesSize()];
            for(int j = 0; j < c.getStatesSize(); j++){
                tempBeliefs[j] = c.getMarginalAt(j);
            }
            beliefs[i] = tempBeliefs;
        }

I can trick the node into thinking it has new evidence.  This was the problem before - adding the likelihood didn't cause the node to think it had new evidence, so JT didn't update the clique potentials correctly.

Josh

I am trying to implement a dynamic bayes net using UnBBayes.  My network topology is as follows:

            d
            |
            |
            V
            c
           / \
          /   \
         <     >
        a       b 

In other words, "a" and "b" both have "c" as a parent, and "c" has "d" as a parent.  The relevant portion of my code follows.

        for(int i = 0; i < NUM_RUNS; i++){
            network.initialize();

            if(i > 0) d.addLikeliHood(beliefs[i-1]);

            a.addFinding(aFindings[i]);
            b.addFinding(bFindings[i]);
            network.updateEvidences();
            float[] tempBeliefs = new float[c.getStatesSize()];
            for(int j = 0; j < c.getStatesSize(); j++){
                tempBeliefs[j] = c.getMarginalAt(j);
            }
            beliefs[i] = tempBeliefs;
        }

The third line of my code is where I believe my problem is.  I am trying to give the net "memory," that is, I want the next state of "c" to be dependent on the previous state.  After inserting evidence into "a" and "b", I query node "c".  The next time around, I insert the posterior probabilities of node "c" into node "d" as likelihood evidence; however, this has no effect on the output of the net (but should).  Regardless of whether or not I comment out that third line, my results don't change.  Adding that likelihood has no effect on the output.  Is UnBBayes unable to handle this type of "soft evidence"?  If so, what is the purpose of the addLikeliHood() method?

This is my first time attempting to use UnBBayes, so there is a high probability that I'm doing something else incorrectly; however, I am certain that the topology and conditional probability tables are set correctly, as my results without the "memory" aspect (i.e. completely removing "d" from the net and code) exactly match the same net implemented with another product.

Any help is greatly appreciated.

Thanks,

Josh

It is a pleasure to finally release a new version. This release has some
major features non-related to MEBN: - Correct classification evaluation; -
Hybrid Bayesian network (Gaussian Mixture) - propagation under development;
- Load and Save XMLBIF version 0.5 (for continuous nodes); - Likelihood
Weighting; - Object-Oriented Bayesian Network (OOBN) - under development; -
Net 3rd Spec for OOBN I/O. The main changes related to MEBN are: - Nested IF
for MEBN CPT; - SSBN algorithm corrections. Hope to receive your feedback.
UnBBayes Team

Dear all,

We announce with pleasure the release of our third MEBN release. Hopefully,
it will be our last BETA release also!

The main change in this version is the algorithm for SSBN construction. It
now considers all evidence below the query node as well.

To make sure this is our last BETA version, we are going to focus in making
JUnit tests and we kindly ask the end user to send us feedback with bugs
and/or new features to be added to rom...@gm....

All the best,
UnBBayes Team!

(From https://sourceforge.net/forum/forum.php?forum_id=848365)


We have finally developed a new Web site (finally in English) and with much
more information for users and developers!

Besides the simplified overview and poker game example, the Web site now has
project information and reports, like Javadoc, Xref for our source code,
changes report, project team, licence, mailing lists, and more!

We will now update the Web site in a weekly basis (at least we hope to!).
So, come and check it out every now and then.

http://unbbayes.sourceforge.net/index.html

All the best,
UnBBayes Team

ERRATA
-------
The following feature was also implemented in this version:
1968915 Pseudocode compiler's optimization
Therefore, it should be 3.4.29-BETA. We will eventually change the zip file
to include this information into the release notes and the about pane.

Thanks!
UnBBayes Team

After you download our evaluation form and fill it out, please send it to
rom...@gm....

Thanks!

UnBBayes Team!

2008/5/20 Rommel Carvalho <rom...@gm...>:

>  We are proud to announce that we have fixed a lot of bugs and finally
> released a new version! UnBBayes-MEBN is now available in version
> 3.3.29-BETA.
>
> We would like to remember all that this is still an beta version, and as
> so, some bugs and limitations are expected.
>
> If you would like to help us or give us some feedbak, please download our
> evaluation form we created to collect data from beta users.
>
> Thank you all for your interest!
>
> UnBBayes Team!
>
>

We are proud to announce that we have fixed a lot of bugs and finally
released a new version! UnBBayes-MEBN is now available in version
3.3.29-BETA.

We would like to remember all that this is still an beta version, and as so,
some bugs and limitations are expected.

If you would like to help us or give us some feedbak, please download our
evaluation form we created to collect data from beta users.

Thank you all for your interest!

UnBBayes Team!

Hi all!

We have just put an evaluation form in our download section so you can help
us with some feedback!

Please, take some time to fill it and send it back to us at
rom...@gm...!

Thank you all for the support!

UnBBayes Team

We are proud to announce that we have finally made it! Now it is for real!
Our first release of UnBBayes-MEBN is available in version 3.0.1-ALPHA.

We would like to remember all that this is an alpha version, and as so, some
bugs and limitations are expected.

If you would like to help us or give us some feedbak, please get in touch
with rom...@gm... and ask for our evaluation form we created to
collect data from alpha and beta users.

Some features:
-Friendly GUI for PR-OWL/MEBN modeling;
-Reasoner for MEBN;
-Pseudo-code for dynamic CPT creation;
-Algorithm for SSBN generation;
-FOL sentences;
-Report of the SSBN generated;
-And more...

Thank you all for your interest!

UnBBayes Team!