opfython.subgraphs

As each type of OPF classifier might need a particular subgraph, we opted to define a unique package to hold them all. The subgraph’s package is in charge of implementing some of its varities, such as the k-nearest neighbour (KNN) subgraph.

A subgraphs package for all common opfython modules.

class opfython.subgraphs.KNNSubgraph(X: Optional[numpy.array] = None, Y: Optional[numpy.array] = None, I: Optional[numpy.array] = None, from_file: Optional[bool] = None)

Bases: opfython.core.Subgraph

A KNNSubgraph is used to implement a k-nearest neightbours subgraph.

__init__(self, X: Optional[numpy.array] = None, Y: Optional[numpy.array] = None, I: Optional[numpy.array] = None, from_file: Optional[bool] = None)

Initialization method.

Parameters

  • X – Array of features.

  • Y – Array of labels.

  • I – Array of indexes.

  • from_file – Whether Subgraph should be directly created from a file.

property best_k(self)

int: Number of adjacent nodes (k-nearest neighbours).

calculate_pdf(self, n_neighbours: int, distance_function: callable, pre_computed_distance: Optional[bool] = False, pre_distances: Optional[numpy.array] = None)

Calculates the probability density function for k neighbours.

Parameters

  • n_neighbours – Number of neighbours in the adjacency relation.

  • distance_function – The distance function to be used to calculate the arcs.

  • pre_computed_distance – Whether OPF should use a pre-computed distance or not.

  • pre_distances – Pre-computed distance matrix.

property constant(self)

float: Constant used to calculate the probability density function.

create_arcs(self, k: int, distance_function: callable, pre_computed_distance: Optional[bool] = False, pre_distances: Optional[numpy.array] = None)

Creates arcs for each node (adjacency relation).

Parameters

  • k – Number of neighbours in the adjacency relation.

  • distance_function – The distance function to be used to calculate the arcs.

  • pre_computed_distance – Whether OPF should use a pre-computed distance or not.

  • pre_distances – Pre-computed distance matrix.

Returns

The maximum possible distances for each value of k.

Return type

(np.array)

property density(self)

float: Density of the subgraph.

eliminate_maxima_height(self, height: float)

Eliminates maxima values in the subgraph that are below the inputted height.

Parameters

height – Height’s threshold.

property max_density(self)

float: Maximum density of the subgraph.

property min_density(self)

float: Minimum density of the subgraph.

property n_clusters(self)

int: Number of assigned clusters.