pyLOM.MANIFOLD

Module contents

pyLOM.MANIFOLD.isomap(X: ndarray, dims: int, n_size: int, comp: int = 1, verbose: bool = True)

Computes Isomap embedding using the algorithm of Tenenbaum, de Silva, and Langford (2000).

Parameters: X : ndarray

NxM Data matrix with N points in the mesh for M simulations

dims: int

Embedding dimensionality to use

n_sizeint

Neighborhood size (number of neighbors for ‘k’ method)

comp: int

Component to embed, if more than 1 (defaults to 1, the largest)

verbose: bool

Display information (default is True)

Returns: Y : ndarray

Contains coordinates for d-dimensional embeddings in Y.

Rlist

Residual variances for the embedding in Y.

Endarray

Edge matrix for neighborhood graph.

pyLOM.MANIFOLD.mds(X: ndarray, dims: int, verbose: bool = True)

Computes the MDS embedding using a custom approach with squared distances and eigen-decomposition.

Parameters: X : ndarray

NxM Data matrix with N points in the mesh for M simulations

dimsint

Embedding dimensionality to use (p in your MATLAB code)

Returns: Y : ndarray

Contains coordinates for d-dimensional embeddings in Y.