This repository contains several methods and datasets used for materials property prediction.
deJong_comp_descriptors.ipynb: Composition based attributes from deJong 2016
Deml_descriptors.ipynb: Formation energy prediction using descriptors from Deml 2016
Deml_matminer.ipynb: Comparison of predictions using the predict_Etot_dHf code (method described in Deml 2016)
Dey_replication.ipynb: Band gap prediction using the method from Dey 2014
Meredig_replication.ipynb: Formation energy prediction from Meredig 2014
Ward_bandgap.ipynb: Band gap prediction using the method from Ward 2016
Ward_energy.ipynb: Formation energy prediction using the method from Ward 2016
Ward_glass_formation.ipynb: Glass formation prediction using the method from Ward 2016
Ward_volume.ipynb: Volume prediction using the method from Ward 2016
bandgap.data: Band gap dataset from Ward 2016
deml_dataset.csv: Formation energy dataset from Deml 2016
deml_predictions.csv: Formation energy predictions generated using deml_dataset.csv and predict_Etot_dHf code
dey_element_data.csv: Element property data used in Dey 2014
dey_training_set.csv: Training set for band gap prediction from Dey 2014
glass.data: Glass formation dataset from Ward 2016
meredig_binary_hull.data: Binary hull training data from Meredig 2014
meredig_full.data: Full training data from Meredig 2014. Contains all data from meredig_binary_hull.data and meredig_stable_ternary.data
meredig_prediction_set.csv: Prediction set from Meredig 2014
meredig_stable_ternary.data: Stable ternary training data from Meredig 2014
oqmd_all.data: Full oqmd dataset from Ward 2016
The run_notebooks.py file creates and runs new notebooks from existing notebooks and datasets.