Abstract:
Random forests are somewhat interpretable, as they return feature values that
can be used to compare the most useful variables for making predictions. Random forests are
generally very accurate and perform well on nonlinear problems with many features, as they
perform implicit feature selection.
