This repository contains model performance on crystal structure classification for binary compounds, derived from 1,400 .cif files using features generated with SAF and CAF.

# Download the repository
git clone https://www.github.com/bobleesj/CAF_SAF_perfomance

# Enter the folder
cd CAF_SAF_perfomance

Install packages listed in requirements.txt:

pip install -r requirements.txt

Or you may install all packages at once:

pip install matplotlib scikit-learn pandas CBFV numpy

Run python main.py

imac@imacs-iMac digitial-discovery % python main.py

Processing outputs/CAF/features_binary.csv with 133 features (1/7).
(1/4) Running SVM model...
(2/4) Running PLS_DA n=2...
(3/4) Running PLS_DA model with the best n...
(4/4) Running XGBoost model...
===========Elapsed time: 8.30 seconds===========

...

Processing outputs/CBFV/oliynyk.csv with 308 features (7/7).
(1/4) Running SVM model...
(2/4) Running PLS_DA n=2...
(3/4) Running PLS_DA model with the best n...
(4/4) Running XGBoost model...
===========Elapsed time: 12.88 seconds===========
imac@imacs-iMac digitial-discovery % 

Check the outputs folder for ML reports, plots, etc.

For Figures 8 and 9, run: python figure_8_case_study.py and python figure_9_case_study.py

Our SAF+CAF features does a great job with classifying crystal structuree for intermetallic binary compouds.

This a PLS-DA Component N=2 result for crystal structure that you can find under outputs/SAF_CAF/PLS_DA_plot

• Compositional features were created using CAF. Ex) outputs/CAF/features_binary.csv
• Structural features were created using SAF Ex) outputs/SAF/binary_features.csv

1. Place a features.csv file in the data folder. It should have a "Structure" column, from which we'll extract all "y" values.
2. Place a CSV file with features in a subdirectory within outputs. Example: outputs/SAF_CAF/binary_features.csv

To automatically format Python code and organize imports:

black -l 79 . && isort .

Run the following command:

python featurizer.py

For help with generating structural data using SAF, contact Bob at sl5400@columbia.edu.