MatFeaLib

MatFeaLib

MatFeaLib (Materials Features Library) is a Python library for generating elemental features from materials composition. These representations are often called “descriptors” and can be used in machine learning and data analysis in Materials Science. To get started you can check the basic usage.

Note

This project is under active development.

It has the following capabilities:

  • Generation of primary atomic features of any given compound, a given list of materials, and a given pandas dataframe

  • Generation of atomic features in statistical form

  • Visualizing the periodic table of elements

MatFeaLib currently includes the following atomic feature collections:

Feature Collection

Description

dft_pbe (dft_pbe_spins)

The DFT calculated atomic properties by using the PBE

approximation, accessible from FHI as the source (source)

dft_hse06 (dft_hse06_spins)

The DFT calculated atomic properties by using the HSE06 hybrid

functional, accessible from FHI as the source (source)

dft_pbe0 (dft_pbe0_spins)

The DFT calculated atomic properties by using the PBE0 hybrid

functional, accessible from FHI as the source (source)

dft_pbesol (dft_pbesol_spins)

The DFT calculated atomic properties by using the PBEsol

approximation, accessible from FHI as the source (source)

dft_pwlda (dft_pwlda_spins)

The DFT calculated atomic properties by using the Local-density

approximation (LDA) parameterized by Perdew and Wang (PW),

accessible from FHI as the source (source)

dft_revpbe (dft_revpbe_spins)

The DFT calculated atomic properties by using the revPBE

approximation, accessible from FHI as the source (source)

lda2015

The DFT calculated atomic properties by using the Local-density

approximation, accessible from the PRL paper as the source

(source)

Mendeleev

The atomic properties from the mendeleev package (source)

Matminer

The atomic properties from the matminer package (source)

Pymatgen

The atomic properties from the pymatgen package (source)

It can also be used by any user-specified atomic feature collection.

The statistical measures of atomic features are useful for multi-stoichiometric material data. The MatFeaLib support the following statistics:

Feature Collection

Description

min

The maximum value

max

The minimum value

sum

The summation of values

mean

The arithmetic mean

std

The standard deviation

var

The unbiased variance

median

The middle value

diff

The difference between maximum and minimum values

gmean

The geometric mean

hmean

The harmonic mean

pmean

The power mean

Kurtosis

The fourth central moment divided by the square of the variance

moment

It is a specific quantitative measure of the shape of a set of points.

expectile

The Expectiles; They are a generalization of the expectation, in the same way as

quantiles are a generalization of the median.

skew

For normally distributed data, the skewness should be about zero. For unimodal

continuous distributions, a skewness value greater than zero means that there

is more weight in the distribution’s right tail.

gstd

The geometric standard deviation; It describes the spread of a set of numbers

where the geometric mean is preferred. It is a multiplicative factor, and so a

dimensionless quantity.

iqr

The interquartile range (IQR); It is the difference between the 75th and 25th

percentile of the data. It measures the dispersion similar to standard deviation

or variance but is much more robust against outliers.

entropy

The Shannon entropy

differential_entropy

The differential entropy

MAD

The median absolute deviation (MAD) computes the median over the absolute

deviations from the median. It is a measure of dispersion similar to the

standard deviation but more robust to outliers.

Get started

Indices and tables