Interpretable matching for causal inference
