mlrl.seco.config module

Author: Michael Rapp (michael.rapp.ml@gmail.com)

Provides utilities that ease the configuration of separate-and-conquer (SeCo) algorithms.

class mlrl.seco.config.HeadTypeParameter

Bases: NominalParameter

A parameter that allows to configure the type of the rule heads that should be used.

HEAD_TYPE_PARTIAL = 'partial'

HEAD_TYPE_SINGLE = 'single-label'

class mlrl.seco.config.HeuristicParameter

Bases: NominalParameter

A parameter that allows to configure the heuristic to be used for learning rules.

class mlrl.seco.config.LiftFunctionParameter

Bases: NominalParameter

A parameter that allows to configure the lift function to be used for the induction of multi-label rules.

LIFT_FUNCTION_KLN = 'kln'

LIFT_FUNCTION_PEAK = 'peak'

OPTION_CURVATURE = 'curvature'

OPTION_K = 'k'

OPTION_MAX_LIFT = 'max_lift'

OPTION_PEAK_LABEL = 'peak_label'

class mlrl.seco.config.PruningHeuristicParameter

Bases: NominalParameter

A parameter that allows to configure the heuristic to be used for pruning individual rules.