One-way analysis for a modelblueprint

Calls one_way() using the modelblueprint's target, exposure, and data slots. Optionally overlays the model's in-sample predictions to produce a lift chart (pass predictions = TRUE).

Usage

# S3 method for class 'modelblueprint'
one_way(
  data,
  var = NA,
  set = c("train", "test", "holdout"),
  predictions = FALSE,
  split = NA_character_,
  bins = 35L,
  type_agg = c("equal_exposure", "equal_range"),
  ret = c("plot", "data"),
  ...,
  precomputed_preds = NULL
)

Arguments

data

A modelblueprint.

var

[character(1)] Feature to plot on the x-axis.

set

[character(1)] Which dataset to use: "train", "test", or "holdout". Default "train".

predictions

[logical(1)] If TRUE, adds in-sample model predictions as a second line (lift chart mode). Default FALSE.

split

[character(1) | NA] Optional split variable.

bins

[integer(1)] Number of bins. Default 35L.

type_agg

[character(1)] "equal_exposure" or "equal_range".

ret

[character(1)] "plot" or "data". Default "plot".

...

Further arguments passed to one_way().

precomputed_preds

[numeric | NULL] Optional vector of pre-computed predictions (one per row of the requested set). Only used when predictions = TRUE. When supplied, the internal predict.modelblueprint() call is skipped.

Value

A plotly object or data.table depending on ret. With multiple variables (e.g. var = NA), a named list with one entry per variable; variables that are skipped with a warning are omitted from the list.