plot_marginals: Plot marginals

In pmcharrison/hvr: Harmonic Viewpoint Regression

Description

Plots the marginal effects for continuous viewpoints in a viewpoint regression model, as extracted by get_marginal.

Usage

plot_marginals(
  model_1,
  model_2 = NULL,
  model_labels = NULL,
  x_lab = "Feature value",
  y_lab = "Effect",
  viewpoint_labels = model_1$viewpoint_labels,
  fill = "blue",
  alpha = 0.25,
  scales = "free",
  ...
)

Arguments

model_1	Viewpoint regression model, as created by viewpoint_regression.
model_2	Optional second viewpoint regression model, to be plotted for comparison.
model_labels	Only relevant if a second viewpoint regression model is provided, in which case this should be a character vector of length 2 corresponding to the plot labels for these two models.
x_lab	(Character scalar) Label for the x axis.
y_lab	(Character scalar) Label for the y axis.
viewpoint_labels	tibble of viewpoint labels.
fill	(Character scalar) Fill colour for the shaded region identifying the 5th-95th percentiles of observed feature values.
alpha	(Numeric scalar) Alpha value for the shaded region.
scales	Passed to facet_wrap.
...	Further arguments to pass to facet_wrap.

Details

The x axis spans the 5th-95th percentiles of theoretically possible feature values as computed in the derivation of the model matrix (see compute_model_matrix). The shaded area identifies the 5th-95th percentiles of observed feature values when computing the model matrix.

pmcharrison/hvr documentation built on April 14, 2020, 2:47 a.m.