Implement and train a deterministic regressor which maps the truth catalog features to the object catalog features. We are looking to overfit the regressor, to confirm that it is complex enough to model the features space and that the scale of the features allow for reasonably easy optimization. Deterministic means that the regressor assumes a given set of truth values only has one possible set of observed values--which is not true! We will address this separately in another issue, using a stochastic regressor model.
Implement and train a deterministic regressor which maps the truth catalog features to the object catalog features. We are looking to overfit the regressor, to confirm that it is complex enough to model the features space and that the scale of the features allow for reasonably easy optimization. Deterministic means that the regressor assumes a given set of truth values only has one possible set of observed values--which is not true! We will address this separately in another issue, using a stochastic regressor model.