riecovest.distance.corr_matrix_distance

riecovest.distance.corr_matrix_distance(A, B)

Computes the correlation matrix distance

0 means that the matrices are equal up to a scaling 1 means that they are maximally different (orthogonal in NxN dimensional space)

Parameters:

• mat1 (np.ndarray of shape (..., N, N)) – First covariance matrix, should be symmetric and positive definite

• mat2 (np.ndarray of shape (..., N, N)) – Second covariance matrix, should be symmetric and positive definite

Returns:

distance – The distance between the two matrices. The shape is the same as the shape of the input matrices, except for the last two dimensions, which are removed. If either of the input matrices are zero, the distance is set to NaN.

Return type:

ndarray of shape (…,)

References

Correlation matrix distaince, a meaningful measure for evaluation of non-stationary MIMO channels - Herdin, Czink, Ozcelik, Bonek