Skip to contents

Output Transformation Log

Source: R/OutputTrafoLog.R
OutputTrafoLog.Rd

Output transformation that takes the logarithm after min-max scaling to \(0, 1\).

See also

Other Output Transformation: OutputTrafo, OutputTrafoStandardize, mlr_output_trafos

Super class

mlr3mbo::OutputTrafo -> OutputTrafoLog

Active bindings

packages

(character())
Set of required packages. A warning is signaled if at least one of the packages is not installed, but loaded (not attached) later on-demand via requireNamespace().

Methods

Public methods

Inherited methods

Method new()

Creates a new instance of this R6 class.

Usage

OutputTrafoLog$new(invert_posterior = FALSE)

Arguments

invert_posterior

(logical(1))
Should the posterior predictive distribution be inverted when used within a SurrogateLearner or SurrogateLearnerCollection? Default is FALSE.

Method update()

Learn the transformation based on observed data and update parameters in $state.

Usage

OutputTrafoLog$update(ydt)

Arguments

ydt

(data.table::data.table())
Data. One row per observation with columns $cols_y.

Method transform()

Perform the transformation.

Usage

OutputTrafoLog$transform(ydt)

Arguments

ydt

(data.table::data.table())
Data. One row per observation with at least columns $cols_y.

Returns

data.table::data.table() with the transformation applied to the columns $cols_y.

Method inverse_transform_posterior()

Perform the inverse transformation on a posterior predictive distribution characterized by the first and second moment.

Usage

OutputTrafoLog$inverse_transform_posterior(pred)

Arguments

pred

(data.table::data.table())
Data. One row per observation characterizing a posterior predictive distribution with the columns mean and se. Can also be a named list of data.table::data.table() with posterior predictive distributions for multiple targets corresponding to (cols_y).

Returns

data.table::data.table() with the inverse transformation applied to the columns mean and se. In the case of the input being a named list of data.table::data.table(), the output will be a named list of data.table::data.table() with the inverse transformation applied to the columns mean and se.

Method inverse_transform()

Perform the inverse transformation.

Usage

OutputTrafoLog$inverse_transform(ydt)

Arguments

ydt

(data.table::data.table())
Data. One row per observation with at least columns $cols_y.

Returns

data.table::data.table() with the inverse transformation applied to the columns $cols_y.

Method clone()

The objects of this class are cloneable with this method.

Usage

OutputTrafoLog$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

if (requireNamespace("mlr3learners") &
    requireNamespace("DiceKriging") &
    requireNamespace("rgenoud")) {
  library(bbotk)
  library(paradox)
  library(mlr3learners)

  fun = function(xs) {
    list(y = xs$x ^ 2)
  }
  domain = ps(x = p_dbl(lower = -10, upper = 10))
  codomain = ps(y = p_dbl(tags = "minimize"))
  objective = ObjectiveRFun$new(fun = fun, domain = domain, codomain = codomain)

  instance = OptimInstanceBatchSingleCrit$new(
    objective = objective,
    terminator = trm("evals", n_evals = 5))

  xdt = generate_design_random(instance$search_space, n = 4)$data

  instance$eval_batch(xdt)

  learner = default_gp()

  output_trafo = ot("log", invert_posterior = TRUE)

  surrogate = srlrn(learner, output_trafo = output_trafo, archive = instance$archive)

  surrogate$update()

  surrogate$output_trafo$state

  surrogate$predict(data.table(x = c(-1, 0, 1)))

  surrogate$output_trafo$invert_posterior = FALSE

  surrogate$predict(data.table(x = c(-1, 0, 1)))
}
#>         mean       se
#>        <num>    <num>
#> 1: -4.202650 1.262598
#> 2: -6.648200 0.848639
#> 3: -5.527866 1.879272