Generator Extraction¶
- class symdet.generator_extraction.generators.GeneratorExtraction(point_cloud: Tensor, delta: float = 0.5, epsilon: float = 0.3, candidate_runs: int = 10)[source]¶
Class to extract generators from point clouds
- point_cloud¶
Point cloud on which to perform regression.
- Type
tf.Tensor
- delta¶
Width of the hyperplane to be considered.
- Type
float
- epsilon¶
Distance between points to be considered as a point pair.
- Type
float
- candidate_runs¶
Number of times to generate candidates before running PCA. Tune if convergence is not found.
- Type
int
- basis¶
Orthonormal basis of the point cloud.
- Type
tf.Tensor
- hyperplane_set¶
Set of all points in the hyperplane.
- Type
tf.Tensor
- point_pairs¶
A list of tuples where each tuple value is an index of a point. The tuple indices correspond to point pairs. i.e. if the tuple is (0, 400), then indices 0 and 400 in the hyperplane set are a pair connected by the application/s of the generators.
- Type
list
- dimension¶
Dimensionality of the points.
- Type
int
- generator_candidates¶
Generator candidates on which to perform PCA.
- Type
list
- constrained_generators¶
Constrained generator candidates.
- Type
np.ndarray:
Methods
perform_generator_extraction([...])Collect all methods and perform the generator extraction.
- perform_generator_extraction(pca_components: int = 4, plot: bool = False, save: bool = False, factor: bool = True, gs_precision: int = 5) Tuple[source]¶
Collect all methods and perform the generator extraction.
- Parameters
pca_components (int) – Number of pca components to checked in the reduction.
plot (bool) – If True, the outcomes will be plotted.
factor (bool) – If true, factor the values by the dimensionality.
save (bool) – If True, and plot is also True, the plots will be saved.
gs_precision (int) – Number of decimals after 0 to consider orthogonal.
- Returns
generators (list) – Return a list of generators.
std_array (list) – explained variance list.