False Positive Rate

Overview

Calculates FP / (FP + TN) for privileged and unprivileged groups. This is the rate of false positives seen out of all the negative events.

Formula

FPR = FP / (FP + TN)

Where:

• TN (True Negatives): The number of negative instances correctly predicted by the model

• FP (False Positives): The number of positive instances incorrectly predicted as positive by the model

Usage

Manually

Python

fpr_priv, fpr_unpriv = false_positive_rate(df, protected_attribute, privileged_group, labels, positive_label, y_true)

print("False Positive Rate for privileged group:", fpr_priv)
print("False Positive Rate for unprivileged group:", fpr_unpriv)

Using Fairness Object

fpr_priv, fpr_unpriv = (fo.compute(false_positive_rate))

Results

False Positive Rate for privileged group: 0.3749999999953125
False Positive Rate for unprivileged group: 0.74999999998125

These results are obtained by using the input data given in the Create Example Data page under Getting Started

Interpretation

A higher FPR indicates a greater likelihood to falsely predict a positive result.