Tool for bias detection, visualization, and mitigation in ML models

1 point

6 years ago

I am a developer of R package fairmodels which facilitates dealing with biased models and ensuring that the model is treating certain subgroups similarly. It is flexible, works with many models (even from different packages) that can be later compared with each other. For a long time, such a tool wasn't available for the R community. I would be more than happy to receive feedback and critique if you want to check it out.

So far there I wrote a blog post on medium and on R-bloggers. R-bloggers: https://www.r-bloggers.com/fairmodels-lets-fight-with-biased-machine-learning-models-part-1%e2%80%8a-%e2%80%8adetection/ Medium : https://medium.com/@ModelOriented/fairmodels-lets-fight-with-biased-machine-learning-models-part-1-detection-6c2786e6c97c

You are also welcome to visit GitHub ( https://github.com/ModelOriented/fairmodels ) and check out vignettes about basic functionalities and advanced bias mitigation techniques which can be found in "articles" section on the website https://modeloriented.github.io/fairmodels/index.html

There are 2 blog posts incoming on those platforms so if you would be interested I can share them later