FairRegBoost: An End-to-End Data Processing Framework for Fair and Scalable Regression

This repository contains the codes needed to reproduce the experiments of our submitted CIKM 2025 Paper

General Information

• For the experiments we reuse work from other GitHubs (put in subfolders of algorithms)
  
• For some approaches, some slight adaptations had to be made to integrate them in our framework, but the general approach was not altered.
  
• The experiments were run under MacOS Sonoma 14.4, Python Version 3.9.6.

HOWTO RUN

• You need to run run_exps_reg.py. The parameters to set for the experiments, like datasets, models to train, are in that file.
  
• This will call the main.py function that runs the experiments, and subsequently calls evaluation_reg.py to evaluate the overall results.
  
• An evaluation file is automatically generated for an experiment, which shows metrics for RMSE and W2, along additional metrics.