Active Learning Distance Base Class

Abstract base class for active-learning distances.

An ALDistance is fitted on a candidate pool or, when the xdt passed to $fit_pool() is NULL, on the search space itself. Concrete distance classes support search-space fitting where the search-space bounds define the scaling or encoding; classes that need empirical pool information throw an error for NULL pools. Reference points are set separately before distances are computed.

Creates a new active-learning distance.

Fit the distance on a finite candidate pool or on the search space.

Set the reference points used by distances().

Compute distances from query points to reference points indexed by i.

Clear all fitted state.

Arguments

id

(character(1))
Identifier of the distance.

packages

(character())
Optional package names required by the distance.

param_set

(paradox::ParamSet)
Configuration parameter set.

label

(character(1))
Label for the object.

man

(character(1))
String in the format [pkg]::[topic] pointing to a manual page.

search_space

(paradox::ParamSet)
Search space describing the candidate-pool columns.

xdt

(NULL | data.table)
Configurations in search-space coordinates. For fit_pool() only, NULL fits the distance to the search space itself where supported.

i

(NULL | integerish())
Optional subset of reference-point indices. NULL uses all reference points.

Value

self.

self.

Numeric matrix with one row per query point and one column per selected reference point.

self.

Fields

label

(character(1)) Label for this object.

man

(character(1)) String in the format [pkg]::[topic] pointing to a manual page.

packages

(character()) Required packages.

is_fitted

(logical(1)) Whether the distance was fitted on a pool or search space.

search_space

(paradox::ParamSet | NULL) Search space used during fitting.

n_pool

(integer(1) | NULL) Number of pool points. NULL if the distance is not fitted or was fitted to the search space without a finite pool.

n_reference_points

(integer(1) | NULL) Number of reference points set for distance computation.

state

(any) Fitted state returned by the subclass.

Dictionary

This class can be retrieved from mlr_al_distances via clx_ald() and clx_alds().

See also

Other ALDistance: ALDistanceGeometry, mlr_al_distances, mlr_al_distances_affine, mlr_al_distances_gower, mlr_al_distances_standardize