This repository accompanies the paper "Quantitative Relaxations for Arrow's Axioms".
The repository contains the following core files:
Voting_rules.py
Implements all complete-ranking voting rules used in the experiments.
fairness_metric.py
Implements:
- The Kendall Tau distance function
- Our quantitative fairness metrics:
- σIIA (Independence of Irrelevant Alternatives)
- σU (Unanimity)
experiments_main.ipynb
VScode / Jupyter notebook to reproduce all key experimental results.
All_Scotland_wards_4th.shp
Shape file for producing the scottish heat map.
- Python
3.10.1 - Votekit
3.1.0
The "scot-elex-main" folder contains the Scottish election dataset.
You can either download the dataset and specify its path in the experiments, or directly set the path to the scot-elex-main folder.
This "scot-elex-main" folder contains the data from Scottish local government elections conducted by ranked voting in the years 2007, 2012, 2017, and 2022. This data was originally assembled and cleaned thanks to the work of David McCune, who collected cast vote records from Scottish sources to create files in a file format called BLT format.