Although practically attractive with high prediction and classification power, complicated learning methods often lack interpretability and reproducibility, limiting their scientific usage. A useful ...