imputeAlgoPerf {aslib}

Imputes algorithm performance for runs which have NA performance values.

Description

The following formula is used for imputation: base +- range.scalar * range.span + N(0, sd = jitter * range.span)
With range.span = max - min.

Returns an object like algo.runs of asscenario, but drops the runstatus and all other measures.

Usage

imputeAlgoPerf(
  asscenario,
  measure,
  base = NULL,
  range.scalar = 0.3,
  jitter = 0,
  impute.zero.vals = FALSE
)

Arguments

asscenario

[ASScenario]
Algorithm selection scenario.

measure

[character(1)]
Measure to impute. Default is first measure in scenario.

base

[numeric(1)]
See formula. Default is NULL, which means maximum of performance values if measure should be minimized, or minimum for maximization case.

range.scalar

[numeric(1)]
See formula. Default is 0.3.

jitter

[numeric(1)]
See formula. Default is 0.

impute.zero.vals

[logical(1)]
Should values which are exactly 0 be imputed to 1e-6? This allows to take the logarithm later on, handy for subsequent visualizations. Note that this really only makes sense for non-negative measures! Default is FALSE.

Value

[data.frame].