MboPlotOptPath: MboPlotOptPath

In PhilippScheller/VisBayesOpt: Visualizing model based optimization runs from package mlrMBO

Description

This class plots the opt state for a specified iteration of the mbo run.

Super class

VisBayesOpt::MboPlot -> MboPlotOptPath

Public fields

stored_pdp

(list())
List for storing already calculated search_space_component effect object for PDPs. Helps to avoid re-calculation in case it was already done.

Methods

Public methods

• MboPlotOptPath$new()

• MboPlotOptPath$plot()

• MboPlotOptPath$clone()

Inherited methods

• VisBayesOpt::MboPlot$set_param_vals()

---

Method new()

Creates a new instance of this R6 class.

Usage

MboPlotOptPath$new(opt_state)

Arguments

opt_state

(OptState).

---

Method plot()

Plots the opt state for a specified iteration of the mbo run.

Usage

MboPlotOptPath$plot(
  highlight_iter = self$param_vals$highlight_iter,
  search_space_component = self$param_vals$search_space_component,
  parallel = TRUE,
  se_factor = 1L
)

Arguments

highlight_iter

(integer(1) | NULL)
Specifies the iteration to be highlighted. The default NULL does not highlight any iteration.

search_space_component

(character(1) | NULL)
Specifies the search_space_component to be calculated in PDP for higher dimensions, i.e. number of features > 2. The default NULL is only a valid input for the 1 dimensional case.

parallel

(boolean(1) | TRUE)
Specifies if computation of PDP is done parallel, i.e. on multiple cores. The default states that multiple cores are used.

se_factor

(numeric(1) | 1)
Specifies the scaling factor for the uncertainty (standard error) estimate for the 1 dimensional case. The default value is that no up/or downscaling is applied.

Returns

(ggplot).

---

Method clone()

The objects of this class are cloneable with this method.

Usage

MboPlotOptPath$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

PhilippScheller/VisBayesOpt documentation built on Sept. 14, 2020, 12:47 p.m.