hyperlearn.linalg
The linalg module contains all mathematical methods, decompositions and mirrors both Numpy’s linalg and Scipy’s linalg modules. HyperLearn’s modules are all optimized and I also showcase some novel new algorithms.
Matrix Decompositions
| Cholesky Decomposition |
cholesky(X, [alpha]) |
:math:`X=UsigmaV^T |
Symmetric Square |
| LU Decomposition |
lu(X, [L_only, U_only, overwrite]) |
X = L @ U |
Any Matrix |
| Singular Value Decomposition |
svd(X, [U_decision, overwrite]) |
X = U * S @ V.T |
Any Matrix |
| QR Decomposition |
qr(X, [Q_only, R_only, overwrite]) |
X = Q @ R |
Any Matrix |
Eigenvalue Problems
| Symmetric EigenDecomposition |
eigh(X, [alpha, svd, overwrite]) |
X = V * L @ V^-1 |
Symmetric Square |
Matrix Inversion
| Cholesky Inverse |
cho_inv(X, [turbo]) |
inv(X) @ X = I |
Symmetric Square |
| Pseudoinverse (Cholesky) |
pinvc(X, [alpha, turbo]) |
inv(X) @ X = I |
Any Matrix |
| Symmetric Pseudoinverse |
pinvh(X, [alpha, turbo, n_jobs]) |
inv(X) @ X = I |
Symmetric Square |
| Pseudoinverse (SVD) |
pinv(X, [alpha, overwrite]) |
inv(X) @ X = I |
Any Matrix |
| Pseudoinverse (LU Decomp) |
pinvl(X, [alpha, turbo, overwrite]) |
inv(X) @ X = I |
Square Matrix |