Plots for Tuning Instances

Visualizations for mlr3tuning::TuningInstanceSingleCrit. The argument type controls what kind of plot is drawn. Possible choices are:

• "marginal" (default): Scatter plots of x versus y. The color of the points shows the batch number.

• "performance": Scatter plots of batch number versus y

• "parameter": Scatter plots of batch number versus input. The color of the points shows the y values.

• "parallel": Parallel coordinates plot. hyperparameters are rescaled by (x - mean(x)) / sd(x).

• "points": Scatter plot of two x dimensions versus. The color of the points shows the y values.

• "surface": Surface plot of two x dimensions versus y values. The y values are interpolated with the supplied mlr3::Learner.

• "pairs": Plots all x and y values against each other.

• "incumbent": Plots the incumbent versus the number of configurations.

Usage

# S3 method for TuningInstanceSingleCrit
autoplot(
  object,
  type = "marginal",
  cols_x = NULL,
  trafo = FALSE,
  learner = mlr3::lrn("regr.ranger"),
  grid_resolution = 100,
  theme = theme_minimal(),
  ...
)

Arguments

object

(mlr3tuning::TuningInstanceSingleCrit.

type

(character(1)):
Type of the plot. See description.

cols_x

(character())
Column names of hyperparameters. By default, all untransformed hyperparameters are plotted. Transformed hyperparameters are prefixed with x_domain_.

trafo

(logical(1))
If FALSE (default), the untransformed hyperparameters are plotted. If TRUE, the transformed hyperparameters are plotted.

learner

(mlr3::Learner)
Regression learner used to interpolate the data of the surface plot.

grid_resolution

(numeric())
Resolution of the surface plot.

theme

(ggplot2::theme())
The ggplot2::theme_minimal() is applied by default to all plots.

...

(ignored).

Value

ggplot2::ggplot().

Examples

if (requireNamespace("mlr3tuning") && requireNamespace("patchwork")) {
  library(mlr3tuning)

  learner = lrn("classif.rpart")
  learner$param_set$values$cp = to_tune(0.001, 0.1)
  learner$param_set$values$minsplit = to_tune(1, 10)

  instance = TuningInstanceSingleCrit$new(
    task = tsk("iris"),
    learner = learner,
    resampling = rsmp("holdout"),
    measure = msr("classif.ce"),
    terminator = trm("evals", n_evals = 10))

  tuner = tnr("random_search")

  tuner$optimize(instance)

  # plot performance versus batch number
  autoplot(instance, type = "performance")

  # plot cp values versus performance
  autoplot(instance, type = "marginal", cols_x = "cp")

  # plot transformed parameter values versus batch number
  autoplot(instance, type = "parameter", trafo = TRUE)

  # plot parallel coordinates plot
  autoplot(instance, type = "parallel")

  # plot pairs
  autoplot(instance, type = "pairs")
}