classification

_______________________________________________
Wekalist mailing list
Wekalist <at> list.scms.waikato.ac.nz
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist

Cotraining

_______________________________________________
Wekalist mailing list
Wekalist <at> list.scms.waikato.ac.nz
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist

Re :classification

_______________________________________________
Wekalist mailing list
Wekalist <at> list.scms.waikato.ac.nz
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist

Re: SVM Problem

> Please I am working with Weka Ver. 3.5.4
> I have a problem using SVM classifier
> I got a message "Problem evaluating classifier: libsvm classes not in  CLASSPATH"
> I want to know how to modify the ini file to set the class path. I have the libsvm.jar file on the same folder
with the weka.jar file.

No need to modify the INI file, as long as you have the libsvm.jar 
listed in your CLASSPATH variable (i.e., the environment variable).

I also recommend upgrading to 3.5.5...
https://list.scms.waikato.ac.nz/mailman/htdig/wekalist/2007-January/008898.html

HTH

Cheers, Peter
--

-- 
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cs.waikato.ac.nz/~fracpete/           Ph. +64 (7) 858-5174

_______________________________________________
Wekalist mailing list
Wekalist <at> list.scms.waikato.ac.nz
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist

Re: classification

>  how to combine more than two classifiers in  a java program,

Just use the meta-classifier "Vote" for instance. For general use of the 
API, have a look here:
   http://weka.sourceforge.net/wiki/index.php/Use_Weka_in_your_Java_code

> and also
> having more than three classes.give me an example java code.

Do you mean "3 class attributes" or "1 class attribute with 3 class 
labels"? The latter is possible with Vote I mentioned aobe, the other is 
not possible within Weka. Weka allows you only to have 1 class attribute 
in the data.

HTH

Cheers, Peter
--

-- 
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cs.waikato.ac.nz/~fracpete/           Ph. +64 (7) 858-5174

Re: Re :classification

> Im looking for the same. help me if u get the code.\"Weka machine learning workbench list.\"  wrote  hello
Sir,      how to combine more than two classifiers in  a java program,and also having more than
three classes.give me an example java code.    with reguards    seema rajaram  Madurai.      

Answer to that post is here:
https://list.scms.waikato.ac.nz/mailman/htdig/wekalist/2007-February/008943.html

Cheers, Peter
--

-- 
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cs.waikato.ac.nz/~fracpete/           Ph. +64 (7) 858-5174

Re: Weighted voting

> Is there a way in Weka to fix the weights that each classifier gets in a
> voting ensemble (e.g. setting 0.48 weight for SMO and 0.26 for both NB and
> Ibk so that in order to override SMO prediction the other two would need to
> agree with one another)? 

AFIK, only by extending Vote yourself, i.e., subclassing Vote and adding 
the necessary code. It does not have that capability built-in.

Cheers, Peter
--

-- 
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cs.waikato.ac.nz/~fracpete/           Ph. +64 (7) 858-5174

StringToWordVector with alternative stopword list

Hi all,

I want to implement my own classifier using Naive Bayes classification and
StringToWordVector.
How do I use StringToWordVector with an alternative stopword list?

Thanks in advance, 
Britta

_______________________________________________
Wekalist mailing list
Wekalist <at> list.scms.waikato.ac.nz
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist

Test Set Problems in Knowledge Flow

Hi,

I'm having problems with the test set component in Knowledge Flow. It 
seems to work when the data is in the same order as that in the test 
file, but not otherwise. There are no error messages when it doesn't work.

I saw that somebody asked a similar question before, but got no answer:-

"1) All classes are represented in the test-set AND
2) The classes are introduced in the test-set in exactly the same order as
they were in the training-set. "

Does anyone have an answer to this? What I really want to do is put a 
new file with the outputs missing through the prediction appender, but 
pretending it is a test set is the only way I've found of doing so. Is 
there a better way? Having no output values seems to upset the test set 
generator.

Thanks
Kevin

--

-- 
The University of Stirling is a university established in Scotland by
charter at Stirling, FK9 4LA.  Privileged/Confidential Information may
be contained in this message.  If you are not the addressee indicated
in this message (or responsible for delivery of the message to such
person), you may not disclose, copy or deliver this message to anyone
and any action taken or omitted to be taken in reliance on it, is
prohibited and may be unlawful.  In such case, you should destroy this
message and kindly notify the sender by reply email.  Please advise
immediately if you or your employer do not consent to Internet email
for messages of this kind.

Re: StringToWordVector with alternative stopword list

> I want to implement my own classifier using Naive Bayes classification and
> StringToWordVector.
> How do I use StringToWordVector with an alternative stopword list?

Lucky you... Just extended the StringToWordVector this week in order to 
be able to use a custom stopword list (option "-stopwords"). Just grab a 
snapshot of Weka from the Weka homepage. More information about the 
structure of the stopword list can be found in the Javadoc of the 
StringToWordVector class.

HTH

Cheers, Peter
--

-- 
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cs.waikato.ac.nz/~fracpete/           Ph. +64 (7) 858-5174