Re: AUC with cost-sensitive classifier, why does it change?

If you use the CostSensitiveClassifier in minimum-expected-cost mode, it
will output "probabilities" that are 0 or 1. It assigns probability 1 to
the class that minimises expected cost and probability 0 to the other
classes.

Cheers,
EIbe

On Fri, Jun 4, 2021 at 8:16 PM David <[email protected]> wrote:

> Hi,
>   it is unclear to me why in a dataset like breast-cancer.arff, using
> RandomForest in a cost-sensitive way (cost of Minimize expected cost =
> True,
> cost matrix=
> 0,10,
> 1,0
> )
> provides me an AUC of *0.498* whereas a standard random forest provides me
> an AUC of *0.634*. Since the cost-sensitive classifier should just apply
> different thresholds on results, and since AUC is threshold-independent, I
> do not understand why the two AUC are different.
>
> Thanks,
> Davide
>
>
>
> --
> Sent from: https://weka.8497.n7.nabble.com/
> _______________________________________________
> Wekalist mailing list -- [email protected]
> Send posts to [email protected]
> To unsubscribe send an email to [email protected]
> 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 -- [email protected]
Send posts to [email protected]
To unsubscribe send an email to [email protected]
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