https://github.com/deutranium/cape-algorithm
Implementation of CAPE algorithm proposed in the paper "Ensuring Fairness under Prior Probability Shifts" by Biswas et al.
https://github.com/deutranium/cape-algorithm
fairness-ml
Last synced: 6 months ago
JSON representation
Implementation of CAPE algorithm proposed in the paper "Ensuring Fairness under Prior Probability Shifts" by Biswas et al.
- Host: GitHub
- URL: https://github.com/deutranium/cape-algorithm
- Owner: deutranium
- License: gpl-3.0
- Created: 2024-01-11T18:09:29.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-01-11T18:27:03.000Z (over 1 year ago)
- Last Synced: 2025-02-09T18:49:21.774Z (8 months ago)
- Topics: fairness-ml
- Language: Jupyter Notebook
- Homepage:
- Size: 26.4 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# CAPE-algorithm
Implementation of CAPE algorithm proposed in the paper ["Ensuring Fairness under Prior Probability Shifts" by Biswas et al.](https://dl.acm.org/doi/abs/10.1145/3461702.3462596), as a part of Fairness, Privacy and Ethics in AI course (Monsoon'22)
### Abstract
(from the paper)
Prior probability shift is a phenomenon where the training and test datasets differ structurally within population subgroups. This phenomenon can be observed in the yearly records of several real-world datasets, for example, recidivism records and medical expenditure surveys. If unaccounted for, such shifts can cause the predictions of a classifier to become unfair towards specific population subgroups. While the fairness notion called Proportional Equality (PE) accounts for such shifts, a procedure to ensure PE-fairness was unknown. In this work, we design an algorithm, called CAPE, that ensures fair classification under such shifts. We introduce a metric, called prevalence difference, which CAPE attempts to minimize in order to achieve fairness under prior probability shifts. We theoretically establish that this metric exhibits several properties that are desirable for a fair classifier. We evaluate the efficacy of CAPE via a thorough empirical evaluation on synthetic datasets. We also compare the performance of CAPE with several state-of-the-art fair classifiers on real-world datasets like COMPAS (criminal risk assessment) and MEPS (medical expenditure panel survey). The results indicate that CAPE ensures a high degree of PE-fairness in its predictions, while performing well on other important metrics.