Decorrelate

Decorrelate is a class to remove correlation from an correlated standard normal vector \(\textbf{z}=[Z_1,...,Z_n]\) with correlation matrix \(\textbf{C}_z=[\rho_{ij}]\). The uncorrelated standard normal vector \(\textbf{u}=[U_1,...,U_n]\) can be calculated as:

\[\textbf{u}^\intercal = \mathbf{H}^{-1} \mathbf{z}^\intercal\]

Decorrelate Class

The Decorrelate class is imported using the following command:

>>> from UQpy.transformations.Decorrelate import Decorrelate

Methods

class Decorrelate(samples_z, corr_z)[source]

A class to remove correlation from correlated standard normal random variables.

Parameters:

samples_z (ndarray) – Correlated standard normal vector of shape (n_samples, n_dimensions).
corr_z (ndarray) – The correlation matrix (\(\mathbf{C_Z}\)) of the standard normal random vector Z .

Attributes

Decorrelate.H: ndarray: The lower diagonal matrix resulting from the Cholesky decomposition of the correlation matrix (\(\mathbf{C_Z}\)).

Decorrelate.samples_u: ndarray: Uncorrelated standard normal vector of shape (n_samples, n_dimensions).