plot_perm_int: Plot permutation-based feature importances

In pmcharrison/hvr: Harmonic Viewpoint Regression

Description

Plots permutation-based feature importances for a viewpoint regression model as extracted from the $perm_int slot.

Usage

plot_perm_int(
  x,
  labels = x$viewpoint_labels,
  axis_label = "Feature importance (bits/chord)",
  order_by_label = FALSE,
  error_bars = FALSE,
  discrete_label = "Discrete",
  continuous_label = "Continuous"
)

## S3 method for class 'viewpoint_regression'
plot_perm_int(
  x,
  labels = x$viewpoint_labels,
  axis_label = "Feature importance (bits/chord)",
  order_by_label = FALSE,
  error_bars = FALSE,
  discrete_label = "Discrete",
  continuous_label = "Continuous"
)

Arguments

x	Viewpoint regression model as created with viewpoint_regression.
labels	tibble of viewpoint labels.
axis_label	(Character scalar) Axis label.
order_by_label	(Logical scalar) If TRUE, bars are ordered by viewpoint label, otherwise they are ordered by quantity.
error_bars	(Logical scalar) Whether or not to plot error bars, which will correspond to the 2.5th/97.5th percentiles of the permutation-importance replicates. These will only be stable for relatively large numbers of replicates (50+).
discrete_label	Label to apply to discrete viewpoints.
continuous_label	Label to apply to continuous viewpoints.

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