Evaluation of Fairness in Machine Learning Models using the UCI Adult Dataset

Abstract

This paper presents a comprehensive analysis of fairness in machine learning models using the UCI Adult Dataset. The study focuses on mitigating biases related to sensitive attributes such as race and gender by reducing the dimensionality of the dataset. We evaluated the performance and fairness of three popular machine learning models—Logistic Regression, Random Forest, and Gradient Boosting—both with and without including sensitive features. The results indicate that while performance metrics remain stable, the fairness metrics reveal significant insights, underscoring the necessity of considering fairness alongside performance in machine learning applications.

Ähnliche Arbeiten