Distance for IDEAL-style active learning methods.
All search-space columns must be numeric or integer. Each geometry column is
affine-scaled to the interval \([-1, 1]\). With a finite pool, scaling uses
the candidate-pool minima and maxima. With xdt = NULL, scaling uses the
finite search-space bounds. Zero-range columns and missing values are mapped
to 0 in the geometry.
Creates an affine active-learning distance.
Format
R6::R6Class object inheriting from ALDistanceGeometry.
See also
Other ALDistance:
ALDistance,
ALDistanceGeometry,
mlr_al_distances,
mlr_al_distances_gower,
mlr_al_distances_standardize