Interpretable classifiers for tabular data via discretization and feature selection