reject_option_classification: Reject option classification

In aif360: Help Detect and Mitigate Bias in Machine Learning Models

Description

Reject option classification is a postprocessing technique that gives favorable outcomes to unpriviliged groups and unfavorable outcomes to priviliged groups in a confidence band around the decision boundary with the highest uncertainty.

Usage

reject_option_classification(
  unprivileged_groups,
  privileged_groups,
  low_class_thresh = 0.01,
  high_class_thresh = 0.99,
  num_class_thresh = as.integer(100),
  num_ROC_margin = as.integer(50),
  metric_name = "Statistical parity difference",
  metric_ub = 0.05,
  metric_lb = -0.05
)

Arguments

unprivileged_groups	A list epresentation for unprivileged group.
privileged_groups	A list representation for privileged group.
low_class_thresh	Smallest classification threshold to use in the optimization. Should be between 0. and 1.
high_class_thresh	Highest classification threshold to use in the optimization. Should be between 0. and 1.
num_class_thresh	Number of classification thresholds between low_class_thresh and high_class_thresh for the optimization search. Should be > 0.
num_ROC_margin	Number of relevant ROC margins to be used in the optimization search. Should be > 0.
metric_name	Name of the metric to use for the optimization. Allowed options are "Statistical parity difference", "Average odds difference", "Equal opportunity difference".
metric_ub	Upper bound of constraint on the metric value
metric_lb	Lower bound of constraint on the metric value

Examples

# Example with Adult Dataset
load_aif360_lib()
ad <- adult_dataset()
p <- list("race",1)
u <- list("race", 0)

col_names <- c(ad$feature_names, "label")
ad_df <- data.frame(ad$features, ad$labels)
colnames(ad_df) <- col_names

lr <- glm(label ~ ., data=ad_df, family=binomial)

ad_prob <- predict(lr, ad_df)
ad_pred <- factor(ifelse(ad_prob> 0.5,1,0))

ad_df_pred <- data.frame(ad_df)
ad_df_pred$label <- as.character(ad_pred)
colnames(ad_df_pred) <- c(ad$feature_names, 'label')

ad_ds <- aif_dataset(ad_df, target_column='label', favor_label = 1,
                     unfavor_label = 0, unprivileged_protected_attribute = 0,
                     privileged_protected_attribute = 1, protected_attribute='race')

ad_ds_pred <- aif_dataset(ad_df_pred, target_column='label', favor_label = 1,
               unfavor_label = 0, unprivileged_protected_attribute = 0,
               privileged_protected_attribute = 1, protected_attribute='race')

roc <- reject_option_classification(unprivileged_groups = u,
                                   privileged_groups = p,
                                   low_class_thresh = 0.01,
                                   high_class_thresh = 0.99,
                                   num_class_thresh = as.integer(100),
                                   num_ROC_margin = as.integer(50),
                                   metric_name = "Statistical parity difference",
                                   metric_ub = 0.05,
                                   metric_lb = -0.05)

roc <- roc$fit(ad_ds, ad_ds_pred)

ds_transformed_pred <- roc$predict(ad_ds_pred)

aif360 documentation built on July 1, 2020, 5:34 p.m.