h2o (version 3.44.0.3)

h2o.inspect_model_fairness: Produce plots and dataframes related to a single model fairness.

Description

Produce plots and dataframes related to a single model fairness.

Usage

h2o.inspect_model_fairness(
  model,
  newdata,
  protected_columns,
  reference,
  favorable_class,
  metrics = c("auc", "aucpr", "f1", "p.value", "selectedRatio", "total"),
  background_frame = NULL
)

Value

H2OExplanation object

Arguments

model

H2O Model Object

newdata

H2OFrame

protected_columns

List of categorical columns that contain sensitive information such as race, gender, age etc.

reference

List of values corresponding to a reference for each protected columns. If set to NULL, it will use the biggest group as the reference.

favorable_class

Positive/favorable outcome class of the response.

metrics

Character vector of metrics to show.

background_frame

Optional frame, that is used as the source of baselines for the marginal SHAP. Setting it enables calculating SHAP in more models but it can be more time and memory consuming.

Examples

if (FALSE) {
library(h2o)
h2o.init()
data <- h2o.importFile(paste0("https://s3.amazonaws.com/h2o-public-test-data/smalldata/",
                              "admissibleml_test/taiwan_credit_card_uci.csv"))
x <- c('LIMIT_BAL', 'AGE', 'PAY_0', 'PAY_2', 'PAY_3', 'PAY_4', 'PAY_5', 'PAY_6', 'BILL_AMT1',
       'BILL_AMT2', 'BILL_AMT3', 'BILL_AMT4', 'BILL_AMT5', 'BILL_AMT6', 'PAY_AMT1', 'PAY_AMT2',
       'PAY_AMT3', 'PAY_AMT4', 'PAY_AMT5', 'PAY_AMT6')
y <- "default payment next month"
protected_columns <- c('SEX', 'EDUCATION')

for (col in c(y, protected_columns))
  data[[col]] <- as.factor(data[[col]])

splits <- h2o.splitFrame(data, 0.8)
train <- splits[[1]]
test <- splits[[2]]
reference <- c(SEX = "1", EDUCATION = "2")  # university educated man
favorable_class <- "0" # no default next month

gbm <- h2o.gbm(x, y, training_frame = train)

h2o.inspect_model_fairness(gbm, test, protected_columns = protected_columns,
                           reference = reference, favorable_class = favorable_class)
}