1 paper
Philippe Goulet Coulombe, Mikael Frenette, Karin Klieber
We reinvigorate maximum likelihood estimation (MLE) for macroeconomic density forecasting through a novel neural network architecture with dedicated mean and variance hemispheres.…