F-Beta Difference

Overview

Calculates the difference in F-Beta scores between privileged and unprivileged groups.

Usage

Manually

Python

# Calculate fb difference

fbd = fb_difference(df, protected_attribute, privileged_group, labels, positive_label, y_true, beta=1)
print(fbd)

Using Fairness Object

result = fo.compute(fb_difference, beta=1)

Results

0.1363636363636364

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

Interpretation

A result of 0 indicates that the privileged and unprivileged groups have identical F-beta scores. A result > 0 indicates that privileged groups have better F-beta scores and a result < 0 indicates unprivileged groups have better F-beta scores.