Ensembling for "deepregression" objects, re-training different random initializations of the same model on the same data and averaging its predictions on the density scale. Ensemble distributions are combined via a tfd_mixture distribution. Includes methods to extract coefficients, the ensemble distribution and fitted values etc. of all ensemble members and docs/tests.
Ensembling for
"deepregression"objects, re-training different random initializations of the same model on the same data and averaging its predictions on the density scale. Ensemble distributions are combined via atfd_mixturedistribution. Includes methods to extract coefficients, the ensemble distribution and fitted values etc. of all ensemble members and docs/tests.