Skip to contents

Plot point of interest with perturbations and predictions

Source: R/plot.R
plot_interest.Rd

This function generates a ggplot visualizing the perturbations of a local model, overlaying the training data and highlighting a specific point of interest (poi). The plot subtitle shows the importances of the first two features.

Usage

plot_interest(kquat)

Arguments

kquat

A list-like object containing at least:

  • local_model$importances numeric vector

  • local_model$glm_predictions numeric vector

  • perturbations data.frame with columns x and y

  • train_data data.frame with columns x, y, and class

  • poi integer or index of the point of interest

Value

A ggplot object

Examples

data(d_vertical)
rfmodel <- randomForest::randomForest(
 class ~ x + y,
 data = d_vertical
)
# Bundle model up
rfmodel_bundled <- bundle::bundle(rfmodel)
ks <- kumquat(
 rfmodel_bundled,
 d_vertical,
  d_vertical[1,],
  class_names = unique(d_vertical$class)
)
#> INFO [2026-06-19 02:26:53] Picking kumquats for row: 1
plot_obj <- plot_interest(ks)