Formal Privacy and Synthetic Data for the American Community Survey

Formal Privacy and Synthetic Data for the American Community Survey
复制标题

美国社区调查的正式隐私和综合数据

DOI:
--
复制
发表时间:
2018
期刊:
影响因子:
--
通讯作者:
Jerome P. Reiter
Jerome P. Reiter
中科院分区:
--
文献类型:
--
作者:
M. Freiman;Rolando A. Rodríguez;Jerome P. Reiter

文献摘要

被引文献

相似文献

美国人口普查局正在扩大正式隐私的使用,以改善避免披露的方法,并使隐私损失的量化。本文讨论了特殊的挑战,正式私人算法应用于美国社区调查(ACS),包括数据的高维加上样本量的限制和使用复杂的调查权重。我们描述了基于模型的方法来创建合成数据的ACS,专注于健康保险的研究。
The U.S. Census Bureau is expanding the use of formal privacy to improve disclosure avoidance methods and to enable quantification of privacy loss. This paper discusses the particular challenges of applying formally private algorithms to the American Community Survey (ACS), including the data’s high dimensionality coupled with sample size limitations and the use of complex survey weights. We describe research on model-based approaches to creating synthetic data for the ACS, focusing on health insurance.