headers
meiyi | 1 Jan 2004 05:16
Picon
Favicon

what does this mean?

Hi,
After executing java weka.classifiers.j48.J48 -t trainingfile.arff -T testfile.arff, it returns an integer (e.g. 5). What does this indicate? out of memory?
 
Any help is greatly appreciated. Thanks!
May Yee
_______________________________________________
Wekalist mailing list
Wekalist <at> list.scms.waikato.ac.nz
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
Permalink | Reply | 0 comments |
headers
Dave Bazell | 1 Jan 2004 21:23
Picon

Output from Neural Network Classifier

The following is a portion of the output I get from the Explorer when using
a neural network classifier.  I assume that the 5:4 means that the actual
class is 5 using a 1-5 class scheme and 4 using a 0-4 class scheme (which I
used in the data).  Can someone confirm this?

Also, I can only get the probability output in the Explorer.  Can this be
added as an option for the command line?

Thanks,

Dave

 inst#,    actual, predicted, error, probability distribution
     1        5:4        4:3      +   0.001  0      0.145 *0.635  0.219
     2        1:0        2:1      +   0     *0.994  0.005  0.001  0
     3        5:4        5:4          0      0      0      0.2   *0.8
     4        3:2        3:2          0      0.152 *0.848  0      0
     5        3:2        3:2          0.001  0.003 *0.996  0      0
     6        3:2        3:2          0.002  0     *0.992  0.006  0
     7        3:2        4:3      +   0.001  0.001  0.322 *0.675  0
     8        3:2        2:1      +   0     *0.642  0.357  0.001  0
     9        2:1        3:2      +   0      0.044 *0.956  0      0
    10        4:3        4:3          0.001  0      0.259 *0.689  0.051
Permalink | Reply | 0 comments |
headers
Himanshu Garg | 2 Jan 2004 08:03
Favicon

Micro Average and Macro Average

Hello,

	Is there a way I by which I can get the Micro Average for a multiple
category classification from a classifier, using Weka.

Thank You
Himanshu.
Permalink | Reply | 0 comments |
headers
Jery Tang | 3 Jan 2004 04:56
Picon

RE: Wekalist Digest, Vol 11, Issue 2

Re Hello,

Of course,

Best Regards
Jery Tang
> -----Original Message-----
> From: wekalist-bounces <at> list.scms.waikato.ac.nz
> [mailto:wekalist-bounces <at> list.scms.waikato.ac.nz] On Behalf Of
> wekalist-request <at> list.scms.waikato.ac.nz
> Sent: 2004年1月3日 7:06
> To: wekalist <at> list.scms.waikato.ac.nz
> Subject: Wekalist Digest, Vol 11, Issue 2
> 
> Send Wekalist mailing list submissions to
> 	wekalist <at> list.scms.waikato.ac.nz
> 
> To subscribe or unsubscribe via the World Wide Web, visit
> 	https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
> or, via email, send a message with subject or body 'help' to
> 	wekalist-request <at> list.scms.waikato.ac.nz
> 
> You can reach the person managing the list at
> 	wekalist-owner <at> list.scms.waikato.ac.nz
> 
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Wekalist digest..."
> 
> 
> Today's Topics:
> 
>    1. Micro Average and Macro Average (Himanshu Garg)
> 
> 
> ----------------------------------------------------------------------
> 
> Message: 1
> Date: 02 Jan 2004 12:33:00 +0530
> From: Himanshu Garg <himanshu <at> students.iiit.net>
> Subject: [Wekalist] Micro Average and Macro Average
> To: weka <wekalist <at> list.scms.waikato.ac.nz>
> Message-ID: <1073026980.18290.1.camel <at> Himanshu>
> Content-Type: text/plain
> 
> Hello,
> 
> 	Is there a way I by which I can get the Micro Average for a multiple
> category classification from a classifier, using Weka.
> 
> Thank You
> Himanshu.
> 
> 
> 
> 
> ------------------------------
> 
> _______________________________________________
> Wekalist mailing list
> Wekalist <at> list.scms.waikato.ac.nz
> https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
> 
> 
> End of Wekalist Digest, Vol 11, Issue 2
> ***************************************
Permalink | Reply | 0 comments |
headers
Mikhail Bilenko | 7 Jan 2004 02:48
Picon

Distributed experiments and RMI

Hello,

I am trying to set up a distributed experiment by following directions in
README_Experiment_Gui.  Instructions for loading RemoteEngine ask to
execute the following:

java -classpath remoteEngine.jar:/path_to_any_jdbc_drivers \
-Djava.security.policy=remote.policy \
-Djava.rmi.server.codebase=file:/path_to_this_directory/remoteEngine.jar \
weka.experiment.RemoteEngine

However, weka.experiment.RemoteEngine is *not* a part of remoteEngine.jar,
so to execute this I have to add the Weka treepath to the classpath
argument.  With that addition, RemoteEngine gets successfully bound in the
RMI registry (I get the message described in README_Experiment_Gui);
however, I get the nasty exception below when I actually try to run the
experiment:

java.rmi.UnmarshalException: error unmarshalling return; nested exception is:
        java.lang.ClassNotFoundException: weka.experiment.RemoteEngine_Stub (no security manager: RMI
class loader disabled)
        java.rmi.UnmarshalException: error unmarshalling return; nested exception is:
        java.lang.ClassNotFoundException: weka.experiment.RemoteEngine_Stub (no security manager: RMI
class loader disabled)
        at sun.rmi.registry.RegistryImpl_Stub.lookup(Unknown Source)
        at java.rmi.Naming.lookup(Naming.java:84)
        at weka.experiment.RemoteExperiment$1.run(RemoteExperiment.java:583)
Caused by: java.lang.ClassNotFoundException: weka.experiment.RemoteEngine_Stub (no security
manager: RMI class loader disabled)

At first I thought that adding the Weka tree path to the classpath is
confusing rmiregistry. I added weka.experiment.RemoteEngine.class (and two
inner class files) to remoteEngine.jar, and ran the command above without
the additional path to Weka.  However, I still get the same error.

Has anyone run into this, or know what could go wrong?  I would greatly
appreciate any advice or help.

The hosts are residing on the same filesystem, and I am using full
hostnames in Experimenter.  I am running Java 1.4.2 on Linux.  I also have
configured the remote.policy file, but I guess this is irrelevant at this
point.

thanks,
Misha
Permalink | Reply | 0 comments |
headers
Hongyu Guo | 7 Jan 2004 23:11
Picon
Favicon

Cross validation and getCurve()

Hi all,

I have the following question when I am using Weka for my Project:

I use the cross validation to do the evaluation:
    Evaluation evaluation = new Evaluation(trainInstances);
    evaluation.crossValidateModel(scheme, testInstances,10);

I don't know how to get the Prediction data from the above evaluation
variable for the getCurve() function to do the ROC analysis. Do I need to run the
getCVPredictions() again to get the Prediction for the getCurve() function?

Could anyone help me out?

Appreciate your help!

Hongyu Guo
Intelligent Data
Permalink | Reply | 0 comments |
headers
Ra Ang | 8 Jan 2004 19:26
Picon

collecting attributes problem

Hi all,

I'd like to use weka Multinomial Bayesian Classifier for text only classification.
Should I gather all possible terms = features in the attributes array?
 Isn't it getting too big when we talk about thousands of docs? 
And even if that method succeeds, while trying to classifiy unknown instances how do I deal with new terms
(unseen so far). I tried classifying an Instance which had one unseen attribute and an
IndexOutOfBoundsException arose.
How can I use Laplace smoothing in weka combined with Bayesian Classifier?

Thanks in advance.

Ralitsa

-----------------------------------------------------------------------
mail.de   +++   mail.de   +++   mail.de   +++   mail.de   +++   mail.de

NEU: Mailen der Spitzenklasse - sichern Sie sich jetzt Ihre kostenlose mail.de-Adresse!
Jetzt kostenlos anmelden unter: http://www.mail.de/?refid=footer

mail.de   +++   mail.de   +++   mail.de   +++   mail.de   +++   mail.de
Permalink | Reply | 0 comments |
headers
Hongyu Guo | 8 Jan 2004 23:25
Picon
Favicon

Cross validation and getCurve()

Hi,

Which class is the default class in the function: getCurve(FastVector predictions)?

Thanks!

Hongyu Guo
---------------------------------
University of Ottawa, Canada
Email:hguo028 <at> site.uottawa.ca
Lab: Intelligent Data(R4026,SITE)
---------------------------------
Permalink | Reply | 0 comments |
headers
Marina Santini | 9 Jan 2004 10:09
Picon

Re: Kappa Statistic

Thank you very much, John, for your reply. It does help.

Regards, Marina


Hinchman, John H (US SSA) wrote:
Message

Marina –

 

Please excuse the delay in responding. I do not know if anyone responded – but here is my interpretation.

 

The Kappa statistic includes measures of class accuracy within an overall measurement of classifier accuracy. It is a better measure of classifier accuracy than overall accuracy because it considers inter-class agreement.

 

Consider the two-class confusion matrix:

 

a          b

------------------

Ta         Fa  | a

Fb        Tb  | b

 

Your equation for K is correct, given that

 

P(A) = (Ta + Tb)/N   (Sum of diagonal in m-class case.)

 

P(E) = (Ta + Fa) x (Ta + Fb)/N2    + (Tb + Fb) x (Tb + Fa)/ N2

 

To generalize P(E) to the m-class case, the terms are Sum_Row_1 x Sum_Col_1 + Sum_Row_2 x Sum_Col_2…

 

One author I found suggests that K < 0.7 is “poor” and that K > 0.8 is “good.”

 

Hope this helps.

 

John Hinchman

BAE SYSTEMS

San Diego, CA

 

 

-----Original Message-----
From: Marina Santini (ITRI) [mailto:Marina.Santini <at> itri.brighton.ac.uk]
Sent: Sunday, November 23, 2003 5:00 AM
To: wekalist <at> list.scms.waikato.ac.nz
Subject: [Wekalist] Kappa Statistic

 

Hallo Everybody,

 

does anyone know if the

Kappa statistic included in the output of

"weka.classifiers.NaiveBayes" (see my

output below) is computed using the following formula:

 

K = P(A) -P(E)

      _________

       1 - P(E)

 

?

 

Thanks a lot for any suggestions, Marina

 

 

^^^^^^^^^^^MyOutput^^^^^^^^^^^^^^^^

=== Stratified cross-validation ===

 

Correctly Classified Instances         132               88      %
Incorrectly Classified Instances        18               12      %
Kappa statistic                          0.8667
Mean absolute error                      0.0246
Root mean squared error                  0.1547
Relative absolute error                 13.6408 %
Root relative squared error             51.4973 %
Total Number of Instances              150    

 


=== Confusion Matrix ===

 

  a  b  c  d  e  f  g  h  i  j   <-- classified as
  9  1  0  4  0  1  0  0  0  0 |  a = interview
  0 15  0  0  0  0  0  0  0  0 |  b = debate
  0  0 15  0  0  0  0  0  0  0 |  c = scripted
  1  0  0 14  0  0  0  0  0  0 |  d = conversation
  0  0  0  0 15  0  0  0  0  0 |  e = advert
  0  0  0  0  0 14  0  1  0  0 |  f = biography
  0  0  0  0  0  0 15  0  0  0 |  g = lore
  0  0  0  0  0  0  0 15  0  0 |  h = reportage
  0  0  0  0  1  0  1  1 10  2 |  i = instructional
  0  1  0  0  0  0  0  1  3 10 |  j = academic


_______________________________________________
Wekalist mailing list
Wekalist <at> list.scms.waikato.ac.nz
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
Permalink | Reply | 0 comments |
headers
Mark Hall | 10 Jan 2004 03:06
Picon
Picon

Re: Weka distributed experiments and RMI

Hi Misha,

Indeed it looks like classes have been accidently omitted from the  
remoteEngine.jar in weka 3.3.6. These are present in the latest  
development version (weka 3.4).

The exception you list below seems to suggest a problem with class  
loading. I assume you have replaced the "path_to_this_directory" with  
the actual path when starting the remote engine.

The other thing that is essential is to specify a codebase when  
starting the Experimenter - this will enable the remoteEngine to  
download weka classes to its virtual machine when it receives a  
RemoteExperiment object via RMI:

java -Djava.rmi.server.codebase=file:/path_to_weka_install/weka.jar  
weka.gui.experiment.Experimenter

See if this helps.

Cheers,
Mark.

On Saturday, January 10, 2004, at 08:16  AM, Mikhail Bilenko wrote:

> Mark, I tried emailing the wekalist with my problem and no-one was  
> able to
> say something about this; I wonder if you happen to know something  
> about
> this.
>
> I am trying to set up a distributed experiment by following directions  
> in
> README_Experiment_Gui.  Instructions for loading RemoteEngine ask to
> execute the following:
>
> java -classpath remoteEngine.jar:/path_to_any_jdbc_drivers \
> -Djava.security.policy=remote.policy \
> -Djava.rmi.server.codebase=file:/path_to_this_directory/ 
> remoteEngine.jar \
> weka.experiment.RemoteEngine
>
> However, weka.experiment.RemoteEngine isn't a part of the  
> remoteEngine.jar
> archive, so to execute this I have to add the Weka treepath to the
> classpath argument.  With that addition, RemoteEngine gets successfully
> bound in the RMI registry (I get the message described in
> README_Experiment_Gui); however, I get the exception below when I  
> actually
> try to run the experiment:
>
> java.rmi.UnmarshalException: error unmarshalling return; nested  
> exception is:
>         java.lang.ClassNotFoundException:  
> weka.experiment.RemoteEngine_Stub (no security manager: RMI class  
> loader disabled)
>         java.rmi.UnmarshalException: error unmarshalling return;  
> nested exception is:
>         java.lang.ClassNotFoundException:  
> weka.experiment.RemoteEngine_Stub (no security manager: RMI class  
> loader disabled)
>         at sun.rmi.registry.RegistryImpl_Stub.lookup(Unknown Source)
>         at java.rmi.Naming.lookup(Naming.java:84)
>         at  
> weka.experiment.RemoteExperiment$1.run(RemoteExperiment.java:583)
> Caused by: java.lang.ClassNotFoundException:  
> weka.experiment.RemoteEngine_Stub (no security manager: RMI class  
> loader disabled)
>
> At first I thought that adding the Weka tree path to the classpath is
> confusing rmiregistry. I added weka.experiment.RemoteEngine.class (and  
> two
> inner class files) to remoteEngine.jar, and ran the command above  
> without
> the additional path to Weka.  However, I still get the same error.
>
> The hosts are residing on the same filesystem, and I am using full
> hostnames in Experimenter.  I am running Java 1.4.2 on Linux.  I also  
> have
> configured the remote.policy file, but I guess this is irrelevant at  
> this
> point.
>
> Would you happen to know if something is wrong with my RemoteEngine
> configuration?  I would greatly appreciate any advice.
>
> thanks,
> Misha
Permalink | Reply | 1 comment |

Gmane