BorgRisk: BorgRisk S4 Class

In BORG: Bounded Outcome Risk Guard for Model Evaluation

BorgRisk S4 Class

Description

Holds the result of borg_inspect or borg_validate: a structured assessment of evaluation risks detected in a workflow or object.

This class stores identified risks, their classification (hard violation vs soft inflation), affected data indices, and recommended remediation actions.

Usage

## S4 method for signature 'BorgRisk'
show(object)

Arguments

object

A BorgRisk object to be printed.

Value

The BorgRisk object, returned invisibly. Called for the side effect of printing a risk assessment summary to the console.

Slots

risks

A list of detected risk objects, each containing:

type

Character string: risk category (e.g., "preprocessing_leak")

severity

Character string: "hard_violation" or "soft_inflation"

description

Character string: human-readable description

affected_indices

Integer vector: row/column indices affected

source_object

Character string: name of the leaky object

n_hard

Integer. Count of hard violations detected.

n_soft

Integer. Count of soft inflation risks detected.

is_valid

Logical. TRUE if no hard violations detected.

train_indices

Integer vector. Row indices in training set.

test_indices

Integer vector. Row indices in test set.

timestamp

POSIXct. When the inspection was performed.

call

Language object. The original call that triggered inspection.

See Also

borg_inspect, borg_validate, borg

Examples

# Create an empty BorgRisk object (no risks detected)
show(new("BorgRisk",
  risks = list(),
  n_hard = 0L,
  n_soft = 0L,
  is_valid = TRUE,
  train_indices = 1:80,
  test_indices = 81:100,
  timestamp = Sys.time(),
  call = quote(borg_inspect(x))
))

BORG documentation built on March 20, 2026, 5:09 p.m.