minimize_cut

tangles.separations.finding.minimize_cut(starting_feature: ndarray, order_derivative: OrderFuncDerivative, max_steps: int = 100000000) → ndarray

Find a locally minimal cut in a graph starting with the cut specified by starting_feature.

Parameters

starting_featurenp.ndarray

-1/1-indicator vector of an initial cut to start the local search with.

order_derivative: OrderFuncDerivative

A function that calculates the discrete derivative (and optionally for better performance the change of the discrete derivative) for finding local minima.

max_stepsint, optional

The maximal number of optimization steps.

Returns

np.ndarray

A -1/1-indicator vector of the found locally minimal cut.