hmer (version 1.6.2)

validation_pairs: Validation Set Diagnostics and Implausibility

Description

Creates pairs plots on the set of validation points of diagnostic suitability and implausibility.

Usage

validation_pairs(ems, points, targets, ranges, nth = 1, cb = FALSE)

Value

A ggplot object.

Arguments

ems

The Emulator object(s).

points

The set of validation points to plot.

targets

The set of targets to match to.

ranges

If provided, this gives the plotting region (see above).

nth

The level of maximum implausibility to plot.

cb

Whether or not the colour scheme should be colourblind friendly.

Details

The plots are organised as follows:

a) Emulated versus simulated output (lower diagonal). This is similar in spirit to comparison_diag: the plotted points are their location in the input space and the points are coloured by the emulator prediction's deviation from the simulator value.

b) Implausibility (upper diagonal). The points are again plotted based on their location in input space, but their colouration is now based on the implausibility of the point.

If ranges is provided, then the plotting region is created relative to these ranges. This can be useful if on later waves of a history match and the plotting is to be done relative to the original input space, rather than the (reduced) parameter space upon which the emulators have been trained.

See Also

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

Examples

 validation_pairs(SIREmulators$ems, SIRSample$validation, SIREmulators$targets)
 wider_ranges <- purrr::map(SIREmulators$ems[[1]]$ranges, ~.*c(-2, 2))
 validation_pairs(SIREmulators$ems, SIRSample$validation,
  SIREmulators$targets, ranges = wider_ranges, cb = TRUE)