[prev in list] [next in list] [prev in thread] [next in thread] 

List:       wekalist
Subject:    Re: [Wekalist] Train and test set are not compatible
From:       Albrecht Zimmermann <tokugawa98 () gmx ! net>
Date:       2004-05-27 13:28:18
Message-ID: 40B5ECF2.9010805 () gmx ! net
[Download RAW message or body]

Ola,

couple of possible problems come to my mind:

1) Supervised discretization for which class information is needed.
2) Different ranges for the numerical values which lets the unsupervised 
classification pick a different splitting point.

                                                Hope it helps, A

Daniela Sandu wrote:

>Dear all,
>
>
>I have been using Weka 3-4 for running data mining
>tests. 
>I have a training set upon which I apply Discretize. 
>The test set is exactly in the
>same format as the train file: the same attribute
>names, the same number of
>attributes, the attributes in the same order, both
>files in the same format
>-arff. I even applied Discretize on the test set.
>But classification is not possible and I receive the
>following message: "Train and test set are not
>compatible". 
>Could anyone please give me some idea what the problem
>might be and what should I do?
> 
>Thanks to all!
>
>
>	
>		
>__________________________________
>Do you Yahoo!?
>Friends.  Fun.  Try the all-new Yahoo! Messenger.
>http://messenger.yahoo.com/ 
>
>_______________________________________________
>Wekalist mailing list
>Wekalist@list.scms.waikato.ac.nz
>https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
>
>
>  
>


_______________________________________________
Wekalist mailing list
Wekalist@list.scms.waikato.ac.nz
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
[prev in list] [next in list] [prev in thread] [next in thread]