The features directory contains all the features used for machine learning. Included in the features directory are the raw GFA data "MDF_DMREF_Metallic_Glasses_v5.csv" before feature processing.
The plots directory contains the figures used throughout the paper and their data in either a json lines or pickle format.
The model directory contains the final XGBoost model fit to build the shapley value summary plot along with the requirements from pip3 freeze. The python version used was Python 3.9.4.
├── README.md
├── features
│ ├── MDF_DMREF_Metallic_Glasses_v5.csv
│ ├── features_dmax_cts.csv
│ ├── features_gfa_cts.csv
│ └── prsd_features_gfa_cts.csv
├── model
│ ├── features_gfa_cts.csv
│ ├── run.py
│ ├── shap.pickle
│ ├── shap.png
│ └── xgboost_model.pickle
└── plots
├── classification
│ ├── original
│ │ ├── PR_positive.png
│ │ └── PR_positive_data.csv
│ └── prsd
│ ├── PRSD_PR_positive_data.csv
│ └── PR_positive_PRSD.png
├── regression
│ ├── parity.jsonl
│ ├── parity.png
│ ├── parity_cv.pickle
│ └── parity_cv.png
├── tgm
│ ├── tgm.jsonl
│ └── tgm.png
├── tprime
│ ├── tprime.jsonl
│ └── tprime.png
├── tstar
│ ├── tstar.jsonl
│ └── tstar.png
├── viscosity
│ ├── plot.py
│ ├── visc_cu50zr50_set_1-run_1_1100K.csv
│ └── visc_cu50zr50_set_1-run_1_1100K.png
└── xgb_select_class
├── shap.pickle
└── shap.png