Extract coefficients from fitted axis-aligned box learner

Extracts fitted subspace parameters from a trained LearnerSubspaceBox object. Returns either explicit hyperparameter bounds or the transformation matrices that map the fitted axis-aligned hyperrectangle to the unit hypercube.

Usage

# S3 method for class 'LearnerSubspaceBox'
coef(object, vectorize = FALSE, ...)

Arguments

object

A LearnerSubspaceBox object with fitted subspace parameters

vectorize

logical whether to return transformation matrices (TRUE) or explicit bounds (FALSE). Default: FALSE

FALSE: Returns data.table with columns hyperparameter, min, max for each hyperparameter
TRUE: Returns data.table with columns hyperparameters (list), A (diagonal matrix), b (translation vector)

...

Additional arguments (currently unused)

Value

A data.table containing fitted subspace parameters.

When vectorize = FALSE (explicit bounds):

hyperparameter: Hyperparameter name
min: Lower bound of fitted interval
max: Upper bound of fitted interval
cat_hp: Categorical level (only if task has categorical hyperparameters)

When vectorize = TRUE (transformation matrices):

hyperparameters: List column containing hyperparameter names
A: List column of diagonal matrices with $1/(max - min)$ on diagonal
b: List column of translation vectors equal to $-min/(max - min)$
cat_hp: Categorical level (only if task has categorical hyperparameters)

Details

For axis-aligned hyperrectangles, the transformation from the fitted subspace to the unit hypercube $[0,1]^d$ is: $$y = Ax + b$$ where:

$A = diag(1/(max - min))$ is a diagonal matrix (independent scaling per dimension)
$b = -min/(max - min)$ is the translation vector
$x \in [min, max]^d$ are original hyperparameter coordinates
$y \in [0,1]^d$ are unit cube coordinates

This maps each hyperparameter from its fitted range $[min, max]$ to $[0, 1]$.

When the task includes categorical hyperparameters, separate coefficient sets are returned for each categorical level, identified by the cat_hp column.

Error Handling

Throws an error if the learner has not been trained. Call train() before extracting coefficients.

Examples

if (FALSE) { # \dontrun{
# Train a box learner
learner <- LearnerSubspaceBox$new(task)
learner$train()

# Get explicit bounds
coef(learner)

# Get transformation matrices
coef(learner, vectorize = TRUE)
} # }