Local Environment-induced Atomic Features (LEAFs)

Local Environment-induced Atomic Features (LEAFs) are derived from comprehensive learning of the local structures of chemical elements in inorganic crystals as described in

A. Vasylenko et. al. "Learning Atoms from Crystal Structures", https://doi.org/10.48550/arXiv.2408.02292

LEAFs have been proven useful as

1. Structurally-informed similarity measure for compositions.
2. Probability measure of elemental substitutions.
3. Featurization for elements and compositions.

We recommend installation via pip

pip install LEAFeatures