search

ModelOriented / fairmodels

Flexible tool for bias detection, visualization, and mitigation
https://fairmodels.drwhy.ai/
GNU General Public License v3.0
85 stars 15 forks source link

Wrong conclusion based upon density plot in basic tutorial vignette #26

Closed FrieseWoudloper closed 3 years ago

FrieseWoudloper commented 3 years ago

In the Basic Tutorial vignette in 1.3.2 it says:

As we can see it is more likely that model will categorize African Americans as not being recidivists than for example Asians.

But the positive label is 'not being a recidivist'. So shouldn't it be the other way around?

As we can see it is more likely that the model will categorize African Americans as being recidivists than for example Asians.

jakwisn commented 3 years ago

Yes, it's right. As we are predicting not being a recidivist the African Americans in that model will be less likely to be classified as non-recidivists.