Comment by d3nt

Broadly speaking, what approach do you use to "build simpler 'explanation models'" from the more complicated "core fraud models"? Do you learn the models separately over the training data, or does the more complicated model somehow influence the training of the simpler model?