This tutorial provides a step-by-step guide on using machine learning to predict the structural properties of materials. We will use a dataset of crystal structures, enhance it with descriptors, and apply machine learning models to predict properties such as shear modulus, elastic anisotropy, and Poisson's ratio.
The tutorial covers the following steps:
- Loading and examining a dataset using pandas.
- Adding descriptors to the dataset using the
matminer
library. - Training and comparing two machine learning methods with
scikit-learn
. - Visualizing the results.
- Python 3.x
- Libraries:
pandas
,matminer
,scikit-learn
,pymatgen
,matplotlib
,seaborn
,IPython
,statsmodels
,tensorflow
The dataset is stored in a JSON file containing structural information and properties of the crystals.