mescla.emma.projection.project_to_subspace

mescla.emma.projection.project_to_subspace(Z, eigenvectors, k)[source]

Orthogonal projection of Z onto the mixing subspace, still in S-space.

Zhat = Z @ V_k @ V_k.T. The difference Z - Zhat is the part of the chemistry that a rank-k mixing model cannot explain – the raw material of the Hooper diagnostics.

Parameters:
Return type:

ndarray