Code collection to fit KRR ML models
Requirements for Example_tutorial.ipynb are:
-
python >= 3.7
-
ase >= 3.19.1
-
numpy >= 1.19.0
-
scikit-learn >= 0.23.1
-
jupyter-notebook >= 6.1.3
-
matplotlib >= 3.3.0
-
QUIP, quippy, GAP https://libatoms.github.io/GAP/installation.html
Molecular database stored in an extended .xyz file. This file serves as
input for the preparation of the kernel matrices.
The molecule info line has to contain an id
and AE
keyword, which is
the id of the molecule in the database and the atomization energy.
Reaction network with bond breaking reactions of types A - > B + C (reaction type I) or A -> B (reaction type II) stored in a .txt file. Each row is a sepatate reaction. The file consists of 4 columns.
Column | Content |
---|---|
1 | Index of molecule A (molecule id in .xyz file) |
2 | Index of molecule B (molecule id in .xyz file) |
3 | Index of molecule C (molecule id in .xyz file), reaction type I or "[x]", reaction type II |
4 | DFT calculated Reaction energy (RE) [eV], RE = sum(AE(products)) - AE(educt) |
Sina Stocker ([email protected])