RE: Problems with running Weka, even after several different installations.

-----Message d'origine-----
De : wekalist-bounces <at> list.scms.waikato.ac.nz
[mailto:wekalist-bounces <at> list.scms.waikato.ac.nz] De la part de Mark Hall
Envoyé : lundi 30 mai 2011 09:17
À : Weka machine learning workbench list.
Objet : Re: [Wekalist] Problems with running Weka,even after several different installations.

On 27/05/11 6:59 PM, CAULIER Yannick (AREVA) wrote:
> -----Message d'origine-----
> De : wekalist-bounces <at> list.scms.waikato.ac.nz 
> [mailto:wekalist-bounces <at> list.scms.waikato.ac.nz] De la part de Mark 
> Hall Envoyé : mercredi 25 mai 2011 09:41 À : Weka machine learning workbench list.
> Objet : Re: [Wekalist] Problems with running Weka,even after several different installations.
>
> On 24/05/11 8:28 PM, CAULIER Yannick (AREVA) wrote:
>> Hello,
>>
>> I have problems with the installation/running of Weka on my actual computer.
>>
>> I have installed several versions of Weka weka-3-6-4jre.exe; 
>> weka-3-6-4.exe; weka-3-6-4jre-x64.exe; weka-3-6-4-x64.exe and also 
>> older versions. I allways encounter following problem after starting
>> Weka:
>>
>> Could not find the main class, program will exit!
>> In a window named: Java Virtual Machine Launcher.
>>
>> I have also tried to deinstalled and install Java: Java Runtime 
>> Environment 6.0 Update 25 without success!

problem installing new weka.attributeSelection algorithm

Hello,

I have developed a new weka.attributeSelection algorithm which I intend 
to use from the weka GUI.
The algorithm works properly from my eclipse java developer GUI, and 
runs correctly tests weka.core.CheckGOE and weka.core.CheckOptionHandler.

Furthermore, when I install the package from the Package Manager and I 
run it from the Classifiy panel in weka Explorer through the 
AttributeSelectedClassifier meta-classifier, my attribute selection 
algorithm works correctly...

but, when I run it from the Attribute Selection panel in weka explorer, 
and I use cross-validation, I get the exception:
"Attributes are not unique!". However if a set "use full training set" I 
get no error.

My attribute selection class extends ASSearch and implements 
OptionHandler. Is there any special method which needs to be implemented 
in order to be able to use the package from the "Attribute Selection" panel?

Thank you,
Pablo.

_______________________________________________
Wekalist mailing list
Send posts to: Wekalist <at> list.scms.waikato.ac.nz
List info and subscription status: https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

Individual classifier citations in Weka?

_______________________________________________
Wekalist mailing list
Send posts to: Wekalist <at> list.scms.waikato.ac.nz
List info and subscription status: https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

Re: Individual classifier citations in Weka?

_______________________________________________
Wekalist mailing list
Send posts to: Wekalist <at> list.scms.waikato.ac.nz
List info and subscription status: https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

Re: Individual classifier citations in Weka?

_______________________________________________
Wekalist mailing list
Send posts to: Wekalist <at> list.scms.waikato.ac.nz
List info and subscription status: https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

Re: Individual classifier citations in Weka?

Double click on the parameter text in explorer/classify (as for
configuring a classifier) and then you have 2 buttons, it is one of
them... don't remember which one.

I hope it can help you. Anyway it is also in the javadoc in weka website.

Best,
Christophe

2011/6/2 Shing-hon Lau <shinghonlau <at> gmail.com>:
> On Wed, Jun 1, 2011 at 15:10, agw10v <agw10v <at> ecs.soton.ac.uk> wrote:
>>
>> Yes, this information is provided for each of the classifiers supported in
>> Weka.
>
> Where can this information be found?  I have looked all over the place, but
> I am unable to find it.
>
> Thanks,
> Shing-hon
>
> _______________________________________________
> Wekalist mailing list
> Send posts to: Wekalist <at> list.scms.waikato.ac.nz
> List info and subscription status:
> https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
> List etiquette:
> http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
>
>

_______________________________________________
Wekalist mailing list
Send posts to: Wekalist <at> list.scms.waikato.ac.nz
List info and subscription status: https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

Difference between NB and MNB

_______________________________________________
Wekalist mailing list
Send posts to: Wekalist <at> list.scms.waikato.ac.nz
List info and subscription status: https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

Percentage split

HI 

I have a question on weka's percentage split functionality. As you know the
percentage split takes the given percentage of the data and trains the
classifier using that percentage. Now let us say we have a list of 100
records to be classified and the percentage split is 40% does weka take the
first 40 records and trains the  classifier on those first 40 records
(records 1 to 40) or else it takes 40 records randomly from the set, i.e.
not necessarily the first 40 records?? 

Thanks for your help.

Terizja
--

-- 
View this message in context: http://old.nabble.com/Percentage-split-tp31756293p31756293.html
Sent from the WEKA mailing list archive at Nabble.com.

_______________________________________________
Wekalist mailing list
Send posts to: Wekalist <at> list.scms.waikato.ac.nz
List info and subscription status: https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

Re: Bayesian Network on continuous features

_______________________________________________
Wekalist mailing list
Send posts to: Wekalist <at> list.scms.waikato.ac.nz
List info and subscription status: https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

Re: Multiclass classification in Weka

On 31/05/11 5:00 PM, Lucian M. Sasu wrote:
> Hello,
>
> Is there a Weka classifier to produce multiclass output? say that I vant
> to solve a regression problem, f : R^m -> R^n, how can I solve it?

Many classifiers in weka handle multi-class problems directly or 
transform multi-class problems into multiple binary class ones (in 
various ways) internally. There is also a MultiClassClassifier meta 
classifier that does this for any binary class classifier using the 
following methods:

1-against-all, random codes, exhaustive code, and 1-against-1.

Cheers,
Mark.

_______________________________________________
Wekalist mailing list
Send posts to: Wekalist <at> list.scms.waikato.ac.nz
List info and subscription status: https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html