MCAConfig

class paminco.algo.mca.MCAConfig(**kwargs)

Settings for MCA algorithms.

Parameters

alphafloat, default=1.01

Relative tolerance for alpha-beta-approximation.

betafloat, default=1

Absolute tolerance for alpha-beta-approximation.

interpolation_accuracyfloat, default=1e-3

# TODO-PW

interpolation_step_size, optional

# TODO-PW

lambda_maxfloat, default=1

Maxmimum parameter (lambda) to find minimum cost flow for.

inverse_methodstr, int or InverseMethod, default=InverseMethod.CHOLESKY

Method to find the inverse Laplacian.

pivot_modeint, str or PivotStepMode, default=PivotStepMode.LEX

Method to choose min edge if len(boundary_edges) > 1.

max_iterint, default=99999

Maximum number of iterations for solver.

printbool, default=False

Whether to print a summary of each iteration.

recomp_intervalint, default=1

If <= 1, recompute inverse Laplacian in every iteration. Else update and recompute every recomp_interval. The higher the value, the less accurate the inverse Laplacian.

round_lambdaint, default=3

Round lambda value for edges to round_lambda after decimal. Used to determine boundary edges.

rounding_margins_baseint, default=-16

Set gamma_dpi with low exponent (in IEEE754) to zero.

rounding_margins_facint, default=-5

Set gamma_dpi with low exponent (in IEEE754) to zero.

__init__(**kwargs)

Methods

__init__(**kwargs)
get_efa_kwargs()
map_kwargs([run])

Attributes

all_options	All available settings for EFA.
inverse_method	Method to calculate inverse Laplacian.
kw_init	Settings for EFA that can be passed only during initialization.
pivot_mode	Method to perform pivot step.