Hi:
I have a question about the preset in CoordinationNumber. Different preset return totally different number in my case. For example, for andalusite if I calculate coordination number of a Si site with preset “VoroniNN” it returns 17, while if I use “MinimumDistanceNN” it returns 4 (which in this case makes more sense). Do we have a general rule when choosing these presets? whats the pros and cons of each preset? Do we have any documents discuss this presets?
Thank you!
Hi Jason
We are preparing a manuscript regarding the different nearneighbor algorithms. it is true they can give wildly different results; nearneighbor finding is less well-defined than you might think.
Based on our tests (which will be presented in the manuscript), I would suggest to use the CrystalNN algorithm in general. We recently pushed a change to CrystalNN in the latest pymatgen that gives even better results if you give oxidation states on the sites in your structure. But even without that, it outperforms the other algorithms in our tests.
As for further documentation available now, you would need to look in the docstrings for the various NN algorithms in pymatgen.analysis.local_env
Best,
Anubhav
Thank you very much Anubhav!
On Wed, Nov 7, 2018, 1:19 PM Anubhav Jain <[email protected] wrote:
Hi Jason
We are preparing a manuscript regarding the different nearneighbor algorithms. it is true they can give wildly different results; nearneighbor finding is less well-defined than you might think.
Based on our tests (which will be presented in the manuscript), I would suggest to use the CrystalNN algorithm in general. We recently pushed a change to CrystalNN in the latest pymatgen that gives even better results if you give oxidation states on the sites in your structure. But even without that, it outperforms the other algorithms in our tests.
As for further documentation available now, you would need to look in the docstrings for the various NN algorithms in pymatgen.analysis.local_env
On Tue, Nov 6, 2018 at 3:32 PM Jason [email protected] wrote:
Hi:
I have a question about the preset in CoordinationNumber. Different preset return totally different number in my case. For example, for andalusite if I calculate coordination number of a Si site with preset “VoroniNN” it returns 17, while if I use “MinimumDistanceNN” it returns 4 (which in this case makes more sense). Do we have a general rule when choosing these presets? whats the pros and cons of each preset? Do we have any documents discuss this presets?
Thank you!
–
You received this message because you are subscribed to the Google Groups “matminer” group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.
–
Best,
Anubhav