NOMAD2018 Transparent Conductive Oxide Materials Predictive Models

Using XGBoost models to predict the bandgap and formation energy of In-Al-Ga-O materials. These models use the following features:

Basic compositional information and unit cell parameters
Elemental properties averaged over the atoms in the cell e.g. electronegativity, ionization potential, orbital radii
Average coordination numbers of the elements
Statistics of M-O bond lengths
Averaged feature vectors obtained from a crystal graph representation of the unit cell

Name		Name	Last commit message	Last commit date
Latest commit History 6 Commits
from_others		from_others
.gitignore		.gitignore
LICENSE.md		LICENSE.md
README.md		README.md
cgnode2vec.py		cgnode2vec.py
dataload.py		dataload.py
exploratory_analysis.ipynb		exploratory_analysis.ipynb
feature_importance.ipynb		feature_importance.ipynb
model_comparison.ipynb		model_comparison.ipynb
process_crystal_graph.ipynb		process_crystal_graph.ipynb
process_xyz_files.ipynb		process_xyz_files.ipynb
properties.py		properties.py
utils.py		utils.py
xgboost_model-with-ext.ipynb		xgboost_model-with-ext.ipynb
xgboost_model.ipynb		xgboost_model.ipynb
xgboost_model_sklearn-with-ext-with-geo-PCA.ipynb		xgboost_model_sklearn-with-ext-with-geo-PCA.ipynb
xgboost_model_sklearn-with-ext-with-geo.ipynb		xgboost_model_sklearn-with-ext-with-geo.ipynb
xgboost_model_sklearn-with-ext.ipynb		xgboost_model_sklearn-with-ext.ipynb

Provide feedback