Data for "Predicting thermal properties of crystals using machine learning"

This project enables the readers of our paper to reproduce our machine learning (ML) results. Below is a summary of the content of the project, as well as the procedure to run the ApplyModels_Phonons.py script.

Description of content

The folder contains the following files:

• ApplyModels_Phonons.py: the main python script to run the ML models on the row training and test data
• X_scalar_id.csv: The full descriptor sets before dimensionality reduction, along with the MaterialsProject ID for each material
• C_v/: The data files and folders for the specific heat capacity ML calculations
  • Figs/: Where the figures will be kept
  • SavedModels/: Where the ML models are saved, and will be loaded
  • SplitDataSets/: Where the 80/20 split training and test sets are saved, and will be loaded
  • X.csv: The CSV file with the dimensionality-reduced descriptors prior to train/test splitting
  • y_C_v.csv: The CSV file that has the values of C_v corresponding to the materials in X.csv prior to splitting
• entropy/: The data files and folders for the entropy ML calculations. It has the same folder structure as C_v
• eps_total_effective/: The data files and folders for the effective polycrystalline dielectric function ML calculations. It has the same folder structure as C_v

Procedure for running the ML code

1. Download the whole content of the folder into a directory, let's call it ML
2. Unzip the three folders C_v.zip, entropy.zip and eps_total_effective.zip
3. Start a shell and cd to the folder ML
4. Type the command python ApplyModels_Phonons.py on your shell

For any questions, you can reach me at [email protected] or [email protected]