Research: Evaluating Fairness-Aware Machine Learning Approaches and Disparate Impact Au...

Research question: How do fairness-aware machine learning models compare to traditional credit-scoring approaches in mitigating disparate impact across demographic groups?

Background & Motivation: Credit scoring algorithms are widely used by financial institutions to assess loan eligibility, yet concerns about algorithmic bias and fairness have grown as these models increasingly influence financial outcomes for individuals. Regulatory bodies and social advocates demand transparency and fairness, especially regarding potential disparate impact on protected groups.

Research Gap / Question: While numerous fairness-aware algorithms have been proposed, there is limited empirical evidence on their effectiveness in real-world credit-scoring applications and how well disparate impact auditing tools can detect and measure bias across demographic groups.

Approach & Expected Contribution: This study will systematically review fairness-aware machine learning methods, implement selected models on publicly available credit datasets, and audit outputs for disparate impact using established fairness metrics. Comparative analysis will quantify improvements and limitations, aiming to inform best practices for fair credit scoring.

Why it Matters: Advancing understanding of fairness in credit-scoring models is crucial for promoting equitable access to financial products, guiding policy, and supporting responsible AI deployment in the financial sector.