About evaluation Metric for regression

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

About evaluation Metric for regression

abasian
Hello to all
I have a question regarding the most appropriate accuracy metric available in weka for regression problems? My paper got rejected recently, and one of the reasons was inappropriate metrics used. I used correlation coefficient CC, Mae, rmse, rrse but still they raised objections. I don't know if there are other options available in weka? Kindly if someone can guide about the better alternative, if there? 
Regards 

_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: 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: About evaluation Metric for regression

Eibe Frank-2
Administrator
The reviewers did not ask for specific metrics then? There are a few other ones for regression available in a couple of WEKA packages:

http://weka.sourceforge.net/packageMetaData/percentageErrorMetrics/index.html
http://weka.sourceforge.net/packageMetaData/logarithmicErrorMetrics/index.html

Cheers,
Eibe

> On 27/09/2019, at 5:35 AM, Asad Ali <[hidden email]> wrote:
>
> Hello to all
> I have a question regarding the most appropriate accuracy metric available in weka for regression problems? My paper got rejected recently, and one of the reasons was inappropriate metrics used. I used correlation coefficient CC, Mae, rmse, rrse but still they raised objections. I don't know if there are other options available in weka? Kindly if someone can guide about the better alternative, if there?
> Regards
> _______________________________________________
> Wekalist mailing list -- [hidden email]
> Send posts to: 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
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: 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: About evaluation Metric for regression

abasian
Hello sir, thank you

The reviewer comment was as

The evaluation criteria that the authors used is very limited. There are
other measures proposed that try to overcome the problems of the MMRE and
others, such as the Standardised Accuracy and several other newer measures.

Best regards



--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: 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: About evaluation Metric for regression

abasian

Is there any way we can use accuracy measures such as Standardised Accuracy
and Pred(x) in weka, directly or in directly? I read a paper from a
well-known researcher that these two measures are unbiased in regression
problems.



--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: 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: About evaluation Metric for regression

Eibe Frank-2
Administrator
Assuming Standardized Accuracy is what is described at


no, that is not implemented in WEKA. The closest related measure in WEKA is probably the relative absolute error. It does not normalise by dividing by the absolute error of a random predictor though: it simply divides by the absolute error obtained by predicting the mean target value in the training data (i.e., what you get using ZeroR as the predictor in WEKA). This tells you at a glance whether your predictor is better than just predicting the mean target value from the training data, in terms of absolute error.

Note that the relative absolute error is a bit of an odd measure because the median is actually the theoretical minimiser of absolute error, not the mean (which is the minimiser of squared error). Thus, it would be more interesting to compare to the median rather than the mean.

Cheers,
Eibe

On Fri, Sep 27, 2019 at 9:22 PM abasian <[hidden email]> wrote:

Is there any way we can use accuracy measures such as Standardised Accuracy
and Pred(x) in weka, directly or in directly? I read a paper from a
well-known researcher that these two measures are unbiased in regression
problems.



--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: 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

_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: 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