print_fairness_object: Print Fairness Object

In fairmodels: Flexible Tool for Bias Detection, Visualization, and Mitigation

Print Fairness Object

Description

Print Fairness Object

Usage

## S3 method for class 'fairness_object'
print(
  x,
  ...,
  colorize = TRUE,
  fairness_metrics = c("ACC", "TPR", "PPV", "FPR", "STP"),
  fair_level = NULL,
  border_width = 1,
  loss_aggregating_function = NULL
)

Arguments

x	fairness_object object
...	other parameters
colorize	logical, whether information about metrics should be in color or not
fairness_metrics	character, vector of metrics. Subset of fairness metrics to be used. The full set is defined as c("ACC", "TPR", "PPV", "FPR", "STP").
fair_level	numerical, amount of fairness metrics that need do be passed in order to call a model fair. Default is 5.
border_width	numerical, width of border between fair and unfair models. If border_width is 1 and model passes one metric less than the fair_level it will be printed with yellow. If border_width is 0 information will be printed in either red or green.
loss_aggregating_function	function, loss aggregating function that may be provided. It takes metric scores as vector and aggregates them to one value. The default is 'Total loss' that measures the total sum of distances to 1. It may be interpreted as sum of bar heights in fairness_check.

Examples

data("german")

y_numeric <- as.numeric(german$Risk) - 1

lm_model <- glm(Risk ~ .,
  data = german,
  family = binomial(link = "logit")
)

rf_model <- ranger::ranger(Risk ~ .,
  data = german,
  probability = TRUE,
  max.depth = 3,
  num.trees = 100,
  seed = 1,
  num.threads = 1
)

explainer_lm <- DALEX::explain(lm_model, data = german[, -1], y = y_numeric)

explainer_rf <- DALEX::explain(rf_model,
  data = german[, -1],
  y = y_numeric
)

fobject <- fairness_check(explainer_lm, explainer_rf,
  protected = german$Sex,
  privileged = "male"
)

print(fobject)

# custom print
print(fobject,
  fairness_metrics = c("ACC", "TPR"), # amount of metrics to be printed
  border_width = 0, # in our case 2/2 will be printed in green and 1/2 in red
  loss_aggregating_function = function(x) sum(abs(x)) + 10
) # custom loss function - takes vector

fairmodels documentation built on Aug. 24, 2022, 1:05 a.m.