Module pca

Module for performing a principle component analysis (PCA).

Functions

numpy rank-2 3Nx3N array, numpy rank-2 MxNx3 array calc_covariance_matrix(coord=None, weights=None)
Calculate the covariance matrix for the structures.

calc_projections(coord=None, num_modes=4)
Calculate the PCA projections.

numpy rank-1 array, numpy rank-3 array, numpy rank2 array pca_analysis(coord=None, weights=None, algorithm='eigen', num_modes=4)
Perform the PCA analysis.

Variables

__package__ = 'lib.structure'

Imports: array, dot, float64, outer, sqrt, zeros, eigh, svd, RelaxError, calc_mean_structure

Function Details

calc_covariance_matrix(coord=None, weights=None)

Calculate the covariance matrix for the structures.

Parameters:

• coord (list of numpy rank-2, Nx3 arrays) - The list of coordinates of all models to superimpose. The first index is the models, the second is the atomic positions, and the third is the xyz coordinates.
• weights (list of float) - The weights for each structure.

Returns: numpy rank-2 3Nx3N array, numpy rank-2 MxNx3 array

The covariance matrix and the deviation matrix.

calc_projections(coord=None, num_modes=4)

Calculate the PCA projections.

Parameters:

• num_modes (int) - The number of PCA modes to calculate.

pca_analysis(coord=None, weights=None, algorithm='eigen', num_modes=4)

Perform the PCA analysis.

Parameters:

• coord (list of numpy rank-2, Nx3 arrays) - The list of coordinates of all models to superimpose. The first index is the models, the second is the atomic positions, and the third is the xyz coordinates.
• weights (list of float) - The weights for each structure.
• algorithm (str) - The PCA algorithm to use (either 'eigen' or 'svd').
• num_modes (int) - The number of PCA modes to calculate.

Returns: numpy rank-1 array, numpy rank-3 array, numpy rank2 array

The PCA values and vectors, and the per structure projections.