Fairness First Clustering: A Multi-Stage Approach for Mitigating Bias

Fair clustering aims to partition a dataset while mitigating bias in the original dataset. Developing fair algorithms has gained increasing attention from machine learning community. In this paper, we propose k-means algorithm, first (FFC), which consists of an initialization stage, relaxation and improvement stage. is employed cluster each group. Then combination step refinement are applied ensure quality guarantee almost fairness. commonly used fairness metric, balance, utilized assess fairness, threshold set allow for improving quality. local search method improve without changing Comparisons carried out between our other state-of-the-art methods on 10 datasets, include both synthetic real-world datasets. The results show that compared with second highest balance value, FFC shares same SSE value one achieves lower values six