返回
Algorithmic Fairness
DOI:10.1146/annurev-financial-110921-125930.png)
摘要
En 中文
This article reviews the recent literature on algorithmic fairness, with a particular emphasis on credit scoring. We discuss human versus machine bias, bias measurement, group versus individual fairness, and a collection of fairness metrics.We then apply these metrics to the US mortgage market, analyzing Home Mortgage Disclosure Act data on mortgage applications between 2009 and 2015.We find evidence of group imbalance in the dataset for both gender and (especially) minority status, which can lead to poorer estimation/prediction for female/minority applicants. Loan applicants are handled mostly fairly across both groups and individuals, though we find that some local male (nonminority) neighbors of otherwise similar rejected female (minority) applicants were granted loans, something that warrants further study. Finally, modern machine learning techniques substantially outperform logistic regression (the industry standard), though at the cost of being substantially harder to explain to denied applicants, regulators, or the courts.
Keyword:
algorithms
machine learning
bias
fairness metrics
credit scoring
期刊
IF:
12
论文数:
3.8K
被引数:
5.5W
机构
引用论文
Dissecting racial bias in an algorithm used to manage the health of populations在用于管理人群健康的算法中剖析种族偏见
SCIENCE
IF45.8
Numerical Investigation Energy Conversion Performance of Tin-Based Perovskite Solar Cells Using Cell Capacitance Simulator
Energies
IF0
Solubility Advantage of Pyrazine-2-carboxamide: Application of Alternative Solvents on the Way to the Future Pharmaceutical DevelopmentPyrazine-2-carboxamide的溶解度优势: 替代溶剂在未来药物开发中的应用

