LowRank.ErrorPropagate

LwRank.ErrorPropagate.py

This module is used to propagate the error of the low rank approximation.

compute_propagated_error(M, P, E)

Compute the propagated error of the low rank approximation. The propagated error Pe is computed using the formula:

Pe = sqrt( (E^2) * (W^2) )

Parameters:

• M (2D array-like) – The matrix used for the low rank approximation.

• P (2D array-like) – The projected matrix.

• E (2D array-like) – The error matrix corresponding to M.

Returns:

Pe – The propagated error matrix corresponding to P.

Return type:

2D array-like