Workflow, data processing, data analysis and predictive ML scripts used for DFT-integrated machine learning methodology in combination with the FLAPW code FLEUR

This code collection contains the Jupyter Notebooks, which were used to implement an integrated machine learning approach into AiiDA workflows, including submission scripts, the integrated machine-learning training, selection and prediction scripts, the processed data in a tabular form, the corresponding analysis, and visualization scripts and the integrated machine learning predictions and metrics themselves for each batch. The methodology has been used on magnetic 2D films with at most three 3d transition metal layers on five FCC noble metal substrate layers.  

The purpose of this publication is to encourage and enable other scientists to implement the method and workflow of integrated machine learning, as described in our upcoming paper, themselves for their respective applications and ab initio codes.

This work was performed as part of the Helmholtz School for Data Science in Life, Earth and Energy (HDS-LEE) and received funding from the Helmholtz Association of German Research Centres.