Skip to contents

Sorted bars, one per person, with the aggregate drawn as a reference line. The people whose bars sit the wrong side of that line are the ones the model is failing – which is the question idiographic work exists to ask.

Usage

plot_subjects(
  x,
  metric = NULL,
  model = NULL,
  scope = NULL,
  n = NULL,
  reference = NULL,
  type = c("points", "density"),
  ...
)

Arguments

x

An idiographic fit.

metric

Metric column. Defaults to rmse, or accuracy when the outcome is a class.

model, scope

Optional filters.

n

Maximum number of people to draw per algorithm.

reference

Optional "pooled" model or "mean" baseline for paired comparisons. Predictions must refer to identical test observations.

type

Person points (default) or densities of paired improvements.

...

Passed to base plotting functions.

Value

Invisibly, the plotted table.

Examples

fit <- fit_ml(srl, y = "effort", x = "efficacy:monitoring", id = "name",
              time = "day", model = "ridge")
plot_subjects(fit)