Replication Package for Fairabel

This artifact is the replication package for the paper “Software Fairness Analysis and Repair via Causal Model-Guided Data Mutation,” accepted by ACM Transactions on Software Engineering and Methodology (TOSEM).

The package contains the source code, processed datasets, and raw experimental results for Fairabel, a causal model-guided data mutation approach for repairing fairness issues in machine learning software.

Contents:
- src/: source code for running Fairabel experiments.
- Dataset/: processed datasets used in the experiments.
- Raw_results/: raw metric outputs from repeated experimental runs.
- README.md: environment setup, execution commands, and result-file format.
- requirements.txt: Python package dependencies.
- LICENSE: software license.

The code supports reproducing Fairabel experiments with logistic regression, random forest, and support vector machine classifiers across the included datasets. The raw result files report predictive performance and fairness metrics, including accuracy, recall, precision, F1 score, MCC, statistical parity difference, average odds difference, and equal opportunity difference.

This package is intended to support future reproducibility and independent validation of the results reported in the paper.