eval_config: Run an algorithm

In pfistfl/rbv2: Reproduce results from randombot-ng

Description

@param learner character
Learner name e.g. "classif.svm". @param task character
OpenML Task ID of the corresponding task. @param configuration list
Named list of hyperparameters. @param parallel integer
Number of cores to use. Set automatically to 1 if not provided by the user. Always set to 1 for xgboost to prevent paralellization errors. @param logfile character
Logfile to write to. @param seed character
Seed to set. Defaults to NULL (no seed). @param extra_metrics list
Additional metrics to compute. List of mlr metrics.

Usage

eval_config(
  learner,
  task_id,
  configuration,
  parallel = NULL,
  logfile = NULL,
  seed = NULL,
  extra_metrics = list(mlr::logloss, mlr::mmce, mlr::auc, mlr::f1, mlr::timetrain,
    mlr::timepredict)
)

Examples

  learner_id = "classif.svm"
  task_id = 3
  configuration = list("gamma" = 0.1, cost  = 10, sample.rate = .1)
  eval_config(learner_id, task_id, configuration)

pfistfl/rbv2 documentation built on Dec. 22, 2021, 7:47 a.m.