Hi everyone, currently working on a machine learning project that aims to use automatic feature extraction (with the XenonPy library) for a specific family of photocatalysts. We didn’t do experiments ourselves and used data and structures from literature published online. However, the chemical space is kinda diverse/messy (idk if this is a good thing). Most of the materials in this realm are composite materials (e.g. Ga/TNT composite), but to make sure that our result is robust we also added in purely organic, inorganic, MOFs and COFs materials.
We planned to use SMILES for organic molecules and pymatgen for for inorganic/MOFs/COFs, but what should we do to the composites??? Really want some suggestions from people who did sth. similar before or used XenonPy library TAT