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