MDL-motivated compression of GLM ensembles increases interpretability and retains predictive power

Open in new window