Quantcast

Questions in Weka

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Questions in Weka

Reham Alamro
Hello,

I have two questions related to the time in Weka, and I would appreciate if you could answer them. These are the questions:

Q1: When we apply any algorithm in Weka, is the time result in general collective (i.e for all instances) or per instance?

Q2: After splitting the whole dataset into training and test sets and then testing (e.g. classifiers) on each one, we get two different time performance measures for each one of them, which are 'Time taken to build model' and 'Time taken to test model'. The question here is what is the difference between these time measures 'Training time' and 'Testing time' for each of training and test sets? 

Your prompt response and help would be greatly appreciated.

Thank you.

Reham

_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Questions in Weka

Eibe Frank-2
Administrator

> On 31/01/2017, at 7:17 AM, Reham Alamro <[hidden email]> wrote:
>
> Q1: When we apply any algorithm in Weka, is the time result in general collective (i.e for all instances) or per instance?

It’s collective, not per instance.

> Q2: After splitting the whole dataset into training and test sets and then testing (e.g. classifiers) on each one, we get two different time performance measures for each one of them, which are 'Time taken to build model' and 'Time taken to test model'. The question here is what is the difference between these time measures 'Training time' and 'Testing time' for each of training and test sets?

Training time is the time required for constructing the classifier (e.g., decision tree) from the training data.

Testing time is the time required to feed the test instances through the classifier, which has already been constructed, to obtain predictions for them.

Cheers,
Eibe

_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Questions in Weka

Reham Alamro
Thank you Eibe for the answer.

About Q2, I would appreciate if you could provide me with more details about the exact difference between these four time measures:

-'Training time' and 'test time' in training set
-'Training time' and 'test time' in test set

Usually, I use training and test time of test set as a final measures, but I want to see if there is any advantage of using the time performance of training set or not.

Thanks again.

Reham

On Tue, Jan 31, 2017 at 5:00 PM, Eibe Frank <[hidden email]> wrote:

> On 31/01/2017, at 7:17 AM, Reham Alamro <[hidden email]> wrote:
>
> Q1: When we apply any algorithm in Weka, is the time result in general collective (i.e for all instances) or per instance?

It’s collective, not per instance.

> Q2: After splitting the whole dataset into training and test sets and then testing (e.g. classifiers) on each one, we get two different time performance measures for each one of them, which are 'Time taken to build model' and 'Time taken to test model'. The question here is what is the difference between these time measures 'Training time' and 'Testing time' for each of training and test sets?

Training time is the time required for constructing the classifier (e.g., decision tree) from the training data.

Testing time is the time required to feed the test instances through the classifier, which has already been constructed, to obtain predictions for them.

Cheers,
Eibe

_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
List info and subscription status: https://list.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: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Questions in Weka

Eibe Frank-2
Administrator
Training time is only applicable to the training set.

Test time is only applicable to the test set (unless the training set is used as the test set).

Cheers,
Eibe

> On 3/02/2017, at 4:05 PM, Reham Alamro <[hidden email]> wrote:
>
> Thank you Eibe for the answer.
>
> About Q2, I would appreciate if you could provide me with more details about the exact difference between these four time measures:
>
> -'Training time' and 'test time' in training set
> -'Training time' and 'test time' in test set
>
> Usually, I use training and test time of test set as a final measures, but I want to see if there is any advantage of using the time performance of training set or not.
>
> Thanks again.
>
> Reham
>
> On Tue, Jan 31, 2017 at 5:00 PM, Eibe Frank <[hidden email]> wrote:
>
> > On 31/01/2017, at 7:17 AM, Reham Alamro <[hidden email]> wrote:
> >
> > Q1: When we apply any algorithm in Weka, is the time result in general collective (i.e for all instances) or per instance?
>
> It’s collective, not per instance.
>
> > Q2: After splitting the whole dataset into training and test sets and then testing (e.g. classifiers) on each one, we get two different time performance measures for each one of them, which are 'Time taken to build model' and 'Time taken to test model'. The question here is what is the difference between these time measures 'Training time' and 'Testing time' for each of training and test sets?
>
> Training time is the time required for constructing the classifier (e.g., decision tree) from the training data.
>
> Testing time is the time required to feed the test instances through the classifier, which has already been constructed, to obtain predictions for them.
>
> Cheers,
> Eibe
>
> _______________________________________________
> Wekalist mailing list
> Send posts to: [hidden email]
> List info and subscription status: https://list.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: [hidden email]
> List info and subscription status: https://list.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: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Loading...