src.geometry.embedding package

Submodules

src.geometry.embedding.laplacian_embedding module

src.geometry.embedding.laplacian_embedding.laplacian_embedding(mat, ncomp, lap_type='geometric', eigen_solver='dense', drop_first=True, in_place=False, **kwargs)

Compute a Laplacian embedding of a graph using either an affinity matrix or a Laplacian matrix.

Parameters:

  • mat (AffinityMatrix | LaplacianMatrix) – The input matrix, either an AffinityMatrix or a LaplacianMatrix.

  • ncomp (int) – Number of components (eigenvectors) to compute.

  • lap_type (Literal['geometric', 'random_walk', 'symmetric']) – Type of Laplacian to construct if mat is an AffinityMatrix. (default: “geometric”).

  • eigen_solver (Literal['dense', 'arpack', 'lobpcg', 'amg']) – Eigenvalue solver to use. (default: “dense”).

  • drop_first (bool) – Whether to drop the first eigenvector (typically trivial). (default: True)

  • in_place (bool) – Whether to modify the input matrix in place. (default: False)

  • kwargs (Any) – Additional keyword arguments passed to the eigen decomposition function.

Returns:

A tuple of eigenvalues and eigenvectors.

Raises:

ValueError – If the input matrix type is not recognized.

Return type:

Tuple[DenseArray, DenseArray]

src.geometry.embedding.ltsa module

Module contents

src.geometry.embedding.laplacian_embedding(mat, ncomp, lap_type='geometric', eigen_solver='dense', drop_first=True, in_place=False, **kwargs)

Compute a Laplacian embedding of a graph using either an affinity matrix or a Laplacian matrix.

Parameters:

  • mat (AffinityMatrix | LaplacianMatrix) – The input matrix, either an AffinityMatrix or a LaplacianMatrix.

  • ncomp (int) – Number of components (eigenvectors) to compute.

  • lap_type (Literal['geometric', 'random_walk', 'symmetric']) – Type of Laplacian to construct if mat is an AffinityMatrix. (default: “geometric”).

  • eigen_solver (Literal['dense', 'arpack', 'lobpcg', 'amg']) – Eigenvalue solver to use. (default: “dense”).

  • drop_first (bool) – Whether to drop the first eigenvector (typically trivial). (default: True)

  • in_place (bool) – Whether to modify the input matrix in place. (default: False)

  • kwargs (Any) – Additional keyword arguments passed to the eigen decomposition function.

Returns:

A tuple of eigenvalues and eigenvectors.

Raises:

ValueError – If the input matrix type is not recognized.

Return type:

Tuple[DenseArray, DenseArray]