Post-Processing

The idea of “post-processing” is relatively simple, as we change neither the training data, nor the model that has been estimated; we simply transform the predictions obtained, to make them “fair” (according to some specific criteria). As actuaries care about calibration, and the associated concept of a “well-balanced” model, quite naturally, we use averages and barycenters. Using optimal transport, we describe techniques, with strong mathematical guarantees, that could be used to get a “fair” pricing model.