International Journal of Science and Research (IJSR)

International Journal of Science and Research (IJSR)
Call for Papers | Fully Refereed | Open Access | Double Blind Peer Reviewed

ISSN: 2319-7064


Downloads: 2 | Views: 89 | Weekly Hits: ⮙2 | Monthly Hits: ⮙2

Research Paper | Computer Science & Engineering | India | Volume 12 Issue 8, August 2023 | Rating: 5.6 / 10


Evaluating Fairness in Healthcare Machine Learning: A Quantitative Approach

K. Roopa


Abstract: As machine learning models become increasingly integral to healthcare, concerns about their fairness in decision - making processes arise. This paper introduces a robust quantitative methodology to measure fairness in healthcare - oriented machine learning algorithms. By evaluating diverse datasets, we identified notable performance disparities across patient subgroups, such as gender and ethnicity. These findings highlight that even models optimized for accuracy can inadvertently perpetuate systemic biases. To counteract these imbalances, we propose specific mitigation strategies, demonstrating their efficacy in enhancing fairness without compromising overall performance. Our research underscores the importance of ensuring equitable AI applications in healthcare, emphasizing that accuracy and fairness must coexist for the optimal benefit of all patients. While there's broad recognition of the need to address fairness, the healthcare domain lacks a comprehensive quantitative metric to assess and counteract it. This paper introduces a novel quantitative measure designed to evaluate fairness in ML algorithms, emphasizing its applicability to healthcare scenarios. We formulate the metric by grounding it in the intricacies of healthcare data and its multifaceted challenges. Our empirical analysis, conducted on multiple healthcare datasets, showcases the utility of our measure in identifying and mitigating biases. The results underscore the metric's potential in aiding the development of more equitable ML models, ensuring that advancements in healthcare ML are both transformative and just for all patient demographics.


Keywords: Algorithmic fairness, machine learning, healthcare, quantitative measure, bias, Adversarial Debiasing


Edition: Volume 12 Issue 8, August 2023,


Pages: 2270 - 2274


How to Download this Article?

Type Your Valid Email Address below to Receive the Article PDF Link


Verification Code will appear in 2 Seconds ... Wait

Top