kcachel/wisefuse

Code and data for "WISE Fusion: Group Fairness Aware Rank Fusion" at CIKM'2024

Within A Group Similiarity Fusion: WISE FUSION

Code and data for "WISE Fuse: Group Fairness Aware Rank Fusion" in Proceedings of the 33rd ACM International Conference on Information and Knowledge Management (CIKM'2024). To reproduce the experiments run adult.py (adult data), econf.py (economic freedom data), worldhappiness.py (world happiness data), ibmhr.py (ibmhr data), disjoint_study.py (overlap credit data), synthetic_tuning.py (dataset for comparing tuning parameter predictability) and synthetic_mallows.py (Mallows base rankings - note the code to generate the Mallows profiles themselves are in data\synthetic-study\generate_mallows.R). Next to produce the plots used in the paper run the script plotting.R in the results/ folder.

All WISE source code is in the src/ folder, and all compared methods are in the comparedmethods/ folder (using code from EPIRA, RAPF, and Fair Queues).

Each dataset is provided in the data/ folder and are derived from publicly released data. However, our repo cannot directly contain the Economic Freedom data. The Fraser institute makes the data publicly available, but users wishing to use it must download it themselves from https://www.fraserinstitute.org/economic-freedom/dataset. Once the efotw-2023-master-index-data-for-researchers-iso.xlsx file is downloaded please place it into the econf/ folder.