headers
Malick.PAYE | 1 Aug 2006 01:00

Malick PAYE/FR/BMEU01 est absent(e).


Je serai absent(e) du  28/07/2006 au 18/08/2006.

Je répondrai à votre message dès mon retour.
Merci,
MP

AVIS : Ce courrier et ses pieces jointes sont destines a leur seul destinataire et peuvent contenir des
informations confidentielles appartenant a bioMerieux. Si vous n'etes pas destinataire, vous etes
informe que toute lecture, divulgation, ou reproduction de ce message et des pieces jointes est
strictement interdite. Si vous avez recu ce message par erreur merci d'en prevenir l'expediteur et de le
detruire, ainsi que ses pieces jointes.

NOTICE: This message and attachments are intended only for the use of their addressee and may contain
confidential information belonging to bioMerieux. If you are not the intended recipient, you are hereby
notified that any reading, dissemination, distribution, or copying of this message, or any attachment,
is strictly prohibited. If you have received this message in error, please notify the original sender
immediately and delete this message, along with any attachments.
Permalink | Reply | 0 comments |
headers
Doorag | 1 Aug 2006 01:50

Market Basket Analysis Questions


When i run the associator apriori, how do i generate rules that have the same
left hand side?

ex. I only want rules A->B that have certain A.  

How have you guys done cross level association rule mining with weka? Is
there a way to classify attributes as parent child?
--

-- 
View this message in context: http://www.nabble.com/Market-Basket-Analysis-Questions-tf2030766.html#a5586241
Sent from the WEKA forum at Nabble.com.
Permalink | Reply | 1 comment |
headers
Peter Reutemann | 1 Aug 2006 02:45
Picon
Picon
Favicon

Re: Market Basket Analysis Questions

> When i run the associator apriori, how do i generate rules that have the same
> left hand side?
> 
> ex. I only want rules A->B that have certain A.  

Are you sure you want the same body in the rules and not the same head?

In Tertius (don't think so that you can do anything like that in 
Apriori) you can set the class index to the actual class attribute of 
the dataset and that classification is to be performed. This setup will 
produce rules with the same head, i.e, the predicted class.

HTH

Cheers, Peter
--

-- 
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cs.waikato.ac.nz/~fracpete/     +64 (7) 838-4466 Ext. 5174
Permalink | Reply | 0 comments |
headers
carola | 1 Aug 2006 10:34
Picon
Picon

class probability for lB1

hi all,
I am experimenting with WEKA and I am wondering how the nearest neighbour
classifier lB1 can output class probabilities because I am using only one
nearest neighbour.
I have 7 classes and I get for each instance 7 probabilities conneted to the
classes.
How can the algorithm get the probabilities?

Thanks,
Carola
Permalink | Reply | 3 comments |
headers
Kimberly Voll | 1 Aug 2006 22:10
Picon
Picon
Favicon

Applying classifier to new data

Hi there,

Forgive me if this is a dumb question, but, if using the GUI interface, 
how do I actually apply the classifier I have trained on new instances 
and, correspondingly, output the results of that application so they 
can be used?  I've poked around the GUI but haven't had any luck.  Do I 
have to use the CLI for the classification of new instances?

This seems obvious, but I cannot find the answer in the documentation 
or mail archives (just noted that several people have asked the same 
question, with no answer :-) ).

Thanks, all!

Cheers,

Kim.

Kimberly D. Voll, PhD
kvoll <at> sfu.ca

"The person who says it cannot be done should not interrupt the person 
doing it"
	- Irene Lewis
Permalink | Reply | 5 comments |
headers
Peter Reutemann | 1 Aug 2006 22:39
Picon
Picon
Favicon

Re:

Please post to the Wekalist...

>  First of all thank you for replying. I tried to read through it, can you give a ".dll" file of
> Weka for .NET framework. 

Like I said earlier, Weka is Java-based. You might be able to generate a 
DLL with some (commercial) native compiler.

http://schmidt.devlib.org/java/native-compilers.html

Cheers, Peter
--

-- 
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cs.waikato.ac.nz/~fracpete/     +64 (7) 838-4466 Ext. 5174
Permalink | Reply | 0 comments |
headers
Peter Reutemann | 2 Aug 2006 02:00
Picon
Picon
Favicon

Re: Applying classifier to new data

> Forgive me if this is a dumb question, but, if using the GUI interface, 
> how do I actually apply the classifier I have trained on new instances 
> and, correspondingly, output the results of that application so they can 
> be used?  I've poked around the GUI but haven't had any luck.  Do I have 
> to use the CLI for the classification of new instances?

It's not possible in the GUI. The Explorer will always re-train your 
model (the only thing that it'll use from a previously trained model, 
are the options).

You will have to save the generated model (right click in the results 
list and choose "Save model") and then use this either from commandline 
or SimpleCLI. E.g., output the predictions for a generated J48 model (in 
a Linux shell, with a configured CLASSPATH environment variable):

   java weka.classifiers.trees.J48 \
     -l /some/where/j48.model \
     -T /some/other/place/dataset.arff \
     -p 0

HTH

Cheers, Peter
--

-- 
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cs.waikato.ac.nz/~fracpete/     +64 (7) 838-4466 Ext. 5174
Permalink | Reply | 4 comments |
headers
Peter Reutemann | 2 Aug 2006 02:06
Picon
Picon
Favicon

Re: class probability for lB1

> I am experimenting with WEKA and I am wondering how the nearest neighbour
> classifier lB1 can output class probabilities because I am using only one
> nearest neighbour.
> I have 7 classes and I get for each instance 7 probabilities conneted to the
> classes.
> How can the algorithm get the probabilities?

It's a "pseudo" distribution: the probability will be 1 for the chosen 
class and 0 for all the other classes. This is due to the implementation 
of the Classifier class: one only needs to implement classifyInstance or 
distributionForInstance, the other one will still return sensible results.

HTH

Cheers, Peter
--

-- 
Peter Reutemann, Dept. of Computer Science, University of Waikato, NZ
http://www.cs.waikato.ac.nz/~fracpete/     +64 (7) 838-4466 Ext. 5174
Permalink | Reply | 2 comments |
headers
Kimberly Voll | 2 Aug 2006 02:06
Picon
Picon
Favicon

Re: Applying classifier to new data

Terrific, thanks Peter.

I'm glad I wasn't crazy and there is in fact no way to do so from the 
GUI. :-)

Cheers,

Kim.

On Aug 1, 2006, at 5:00 PM, Peter Reutemann wrote:

>> Forgive me if this is a dumb question, but, if using the GUI 
>> interface, how do I actually apply the classifier I have trained on 
>> new instances and, correspondingly, output the results of that 
>> application so they can be used?  I've poked around the GUI but 
>> haven't had any luck.  Do I have to use the CLI for the 
>> classification of new instances?
>
> It's not possible in the GUI. The Explorer will always re-train your 
> model (the only thing that it'll use from a previously trained model, 
> are the options).
>
> You will have to save the generated model (right click in the results 
> list and choose "Save model") and then use this either from 
> commandline or SimpleCLI. E.g., output the predictions for a generated 
> J48 model (in a Linux shell, with a configured CLASSPATH environment 
> variable):
>
>   java weka.classifiers.trees.J48 \
>     -l /some/where/j48.model \
(Continue reading)

Permalink | Reply | 3 comments |
headers
bernhard | 2 Aug 2006 04:31
Picon
Picon

Re: Applying classifier to new data

Hi,

not sure whether you will appreciate this then or not,
but there is a way to do what you want in the Explorer:

If you right-click on the result list you'll get a popup
with options including "Load model" which allows you to load
some saved model; and there is another option "Re-evaluate
model on the current test-set"; obviously before you can
do that, you will have to load some test-set: tick "Supplied
test set" and then the "Set" button and select your file ...

Bernhard

Kimberly Voll wrote:
> Terrific, thanks Peter.
> 
> I'm glad I wasn't crazy and there is in fact no way to do so from the
> GUI. :-)
> 
> 
> Cheers,
> 
> Kim.
> 
> On Aug 1, 2006, at 5:00 PM, Peter Reutemann wrote:
> 
>>> Forgive me if this is a dumb question, but, if using the GUI
>>> interface, how do I actually apply the classifier I have trained on
>>> new instances and, correspondingly, output the results of that
(Continue reading)

Permalink | Reply | 2 comments |

Gmane