behaviour_plot: Output Plotting

In hmer: History Matching and Emulation Package

Output Plotting

Description

A simple diagnostic plot that compares the output values to input values, for each possible combination. If emulators are provided, the emulator predictions are plotted; otherwise the model outputs are plotted.

Usage

behaviour_plot(
  ems,
  points,
  model = missing(ems),
  out_names = unique(names(collect_emulators(ems))),
  targets = NULL
)

Arguments

ems	A set of Emulator objects.
points	A set of points at which to evaluate the emulator expectation
model	If TRUE, use the model outputs; else use emulator expectation
out_names	If no emulators are provided, use this argument to indicate outputs.
targets	If targets are provided, these are added into the plots.

Details

If emulators are provided, then the points argument is optional: if given then the emulator predictions will correspond to those at the points provided. If no points are provided, 100*d (where d is the number of input parameters) are sampled uniformly from the space and used to predict at.

If no emulators are provided, then points must be provided, along with the names of the outputs to plot; each named output must exist as a column in the points data.frame.

Value

The dependency plots.

See Also

Other visualisation tools: diagnostic_wrap(), effect_strength(), emulator_plot(), hit_by_wave(), output_plot(), plot_actives(), plot_lattice(), plot_wrap(), simulator_plot(), space_removed(), validation_pairs(), wave_dependencies(), wave_points(), wave_values()

Examples

 behaviour_plot(SIREmulators$ems, model = FALSE)
 behaviour_plot(points = SIRSample$training, out_names = names(SIREmulators$ems))
 #> Throws a warning
 behaviour_plot(SIRMultiWaveEmulators, model = TRUE, targets = SIREmulators$targets)

hmer documentation built on June 22, 2024, 9:22 a.m.