Use trained classifier on a new/separate dataset

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Use trained classifier on a new/separate dataset

Carlos G. Gallo
Hello,
I have trained a random tree forest with 40 cross-validation fold. The
results are encouraging (Kappa  .76). I now want to apply this trained
classifier to a new different dataset.
The problem I have is that when the predictions are made, I don't find
a way to link the output to the original dataset IDs.  I saw some
postings that report the predictions do not come out in the same order
as the input.

How can I link the predicted output of the classifier to the original ID?
Thank you,
Andres
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to [hidden email]
To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|

Re: Use trained classifier on a new/separate dataset

Peter Reutemann-3
If you are applying a trained model to another dataset, then the order will stay the same.

When performing cross-validation or random splits then the order will be different to the input, due to the randomization.

For instance IDs see here:
https://waikato.github.io/weka-wiki/instance_id/

Cheers, Peter

On November 17, 2020 9:42:19 AM GMT+13:00, "Carlos G. Gallo" <[hidden email]> wrote:

>Hello,
>I have trained a random tree forest with 40 cross-validation fold. The
>results are encouraging (Kappa  .76). I now want to apply this trained
>classifier to a new different dataset.
>The problem I have is that when the predictions are made, I don't find
>a way to link the output to the original dataset IDs.  I saw some
>postings that report the predictions do not come out in the same order
>as the input.
>
>How can I link the predicted output of the classifier to the original
>ID?
>Thank you,
>Andres
>_______________________________________________
>Wekalist mailing list -- [hidden email]
>Send posts to [hidden email]
>To unsubscribe send an email to [hidden email]
>To subscribe, unsubscribe, etc., visit
>https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
>List etiquette:
>http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

--
Peter Reutemann
Dept. of Computer Science
University of Waikato, NZ
+64 (7) 577-5304
http://www.cms.waikato.ac.nz/~fracpete/
http://www.data-mining.co.nz/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to [hidden email]
To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|

Re: Use trained classifier on a new/separate dataset

Bill Bane
You can also, on the Classify panel, select More Options and then under
Output Predictions enter the attributes to output with the predictions.
Often you will have an ID# or other identifying attribute that you can
thereby link to each prediction



--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to [hidden email]
To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html