Creating Unbiased Machine Learning Models by Design

Unintended bias against protected groups has become a key obstacle to the widespread adoption of machine learning methods. This work presents modeling procedure that carefully builds models around class information in order make sure final model is independent status, even nonlinear sense. works for any method. The was tested on subprime credit card data combined with demographic by zip code from US Census. census serves as an imperfect proxy borrower demographics but illustrate procedure.