mlr3 (version 1.0.0)

uhashes: Obtain specific uhashes from a BenchmarkResult

Description

In a BenchmarkResult, each ResampleResult is uniquely identified by a hash (uhash). Operations that select specific ResampleResults from a BenchmarkResult operate using these hashes. This function allows to obtain uhashes for specific learners, tasks, and resamplings.

If you want more control, you can also directly obtain the uhash table from the BenchmarkResult via the field $uhash_table.

Usage

uhashes(bmr, learner_ids = NULL, task_ids = NULL, resampling_ids = NULL)
uhash(bmr, learner_id = NULL, task_id = NULL, resampling_id = NULL)

Arguments

bmr

(BenchmarkResult)
Benchmark result.

learner_ids

(character() | NULL)
Learner IDs.

task_ids

(character() | NULL)
Task IDs.

resampling_ids

(character() | NULL)
Resampling IDs.

learner_id

(character(1) | NULL)
Learner ID.

task_id

(character(1) | NULL)
Task ID.

resampling_id

(character(1) | NULL)
Resampling ID.

Examples

design = benchmark_grid(
  tsks(c("sonar", "iris")),
  lrns(c("classif.debug", "classif.featureless", "classif.rpart")),
  rsmp("holdout")
)
bmr = benchmark(design)
bmr
bmr$uhashes
uhash(bmr, learner_id = "classif.debug", task_id = "sonar", resampling_id = "holdout")
uhashes(bmr, learner_ids = c("classif.debug", "classif.featureless"))