I am wondering that a brute-force way of generating and fitting a set of clusters within a given cutoff radius might not be good for a 2D cluster expansion, where I only care about the interactions between adsorbates.
Is there a way to exclude clusters that contain atoms of the substrate?
Do you think this selection of clusters will be helpful in terms of physical interpretation and fitting robustness?
Dear CASM developers,
I am wondering that a brute-force way of generating and fitting a set of clusters within a given cutoff radius might not be good for a 2D cluster expansion, where I only care about the interactions between adsorbates.
Is there a way to exclude clusters that contain atoms of the substrate?
Do you think this selection of clusters will be helpful in terms of physical interpretation and fitting robustness?
Thanks, Lei