一个共识的隐私指标框架,用于合成数据
Lisa Pilgram1,2,3, Fida Kamal Dankar2, Jörg Drechsler4,5,6
1School of Epidemiology and Public Health, University of Ottawa, Ottawa, ON K1H 8M5, Canada.
Patterns (New York, N.Y.)
|October 27, 2025
概括
合成数据生成提供了一种分享敏感信息的方法,但必须证明隐私. 本研究提出了一个评估隐私风险的框架,建议不要使用相似度指标,并专注于成员身份和属性披露,以获得更好的保护.
科学领域:
- 计算机科学 计算机科学
- 数据 隐私 数据 隐私 数据
- 统计建模 统计建模
背景情况:
- 合成数据生成对于敏感部门的二次数据使用至关重要.
- 确保合成数据中的个人隐私对于道德和法律合规至关重要.
- 现有的隐私评估指标往往缺乏明确的解释.
研究的目的:
- 开发一个基于共识的框架来评估合成数据中的隐私.
- 识别和推评估隐私风险的有效指标.
- 引导负责任地采用合成数据生成.
主要方法:
- 专家达成共识的过程,以制定隐私评估框架.
- 分析常用的隐私指标及其局限性.
- 确定关键的披露类型:成员资格和属性披露.
主要成果:
- 通常使用的真实数据和合成数据之间的相似度指标缺乏准确的解释,应避免用于隐私评估.
- 在评估会员资格和属性披露的重要性方面达成共识.
- 拟议的框架为有效衡量这些披露提供了建议.
结论:
- 为合成数据隐私评估建立了一个强大的框架,优先考虑会员资格和属性披露.
- 该框架的建议适用于不同的私有合成数据.
- 需要进一步的研究来支持广泛采用保护隐私的合成数据.
相关概念视频
Statistical Analysis: Overview
14.1K
When we take repeated measurements on the same or replicated samples, we will observe inconsistencies in the magnitude. These inconsistencies are called errors. To categorize and characterize these results and their errors, the researcher can use statistical analysis to determine the quality of the measurements and/or suitability of the methods.
One of the most commonly used statistical quantifiers is the mean, which is the ratio between the sum of the numerical values of all results and the...
One of the most commonly used statistical quantifiers is the mean, which is the ratio between the sum of the numerical values of all results and the...
14.1K
Censoring Survival Data
518
Survival analysis is a statistical method used to analyze time-to-event data, often employed in fields such as medicine, engineering, and social sciences. One of the key challenges in survival analysis is dealing with incomplete data, a phenomenon known as "censoring." Censoring occurs when the event of interest (such as death, relapse, or system failure) has not occurred for some individuals by the end of the study period or is otherwise unobservable, and it might have many different...
518
Empirical Method to Interpret Standard Deviation
9.3K
The empirical rule, also known as the three-sigma rule, allows a statistician to interpret the standard deviation in a normally distributed dataset. The rule states that 68% of the data lies within one standard deviation from the mean, 95% lies within two standard deviations from the mean, and 99.7% lies within three standard deviations from the mean. Additionally, this rule is also called the 68-95-99.7 rule.
This rule is used widely in statistics to calculate the proportion of data values...
This rule is used widely in statistics to calculate the proportion of data values...
9.3K
Sign Test for Median of Single Population
339
In general, the sign test serves as a nonparametric method to test hypotheses about the median of a single population when the data does not follow a known distribution. This simplicity makes it particularly useful for small sample sizes or when the assumptions of parametric tests cannot be met. The process begins with identifying a null hypothesis, typically stating that the population median equals a specific value. The alternative hypothesis could be that the median is either not equal to,...
339
Confidence Coefficient
10.4K
The confidence coefficient is also known as the confidence level or degree of confidence. It is the percent expression for the probability, 1-α, that the confidence interval contains the true population parameter assuming that the confidence interval is obtained after sufficient unbiased sampling; for example, if the CL = 90%, then in 90 out of 100 samples the interval estimate will enclose the true population parameter. Here α is the area under the curve, distributed equally under...
10.4K
Wilcoxon Signed-Ranks Test for Median of Single Population
430
The Wilcoxon signed-rank test for the median of a single population is a nonparametric test used to evaluate whether the median of a population differs from a specified value. Unlike parametric tests, it does not require data to follow a normal distribution, making it suitable for non-normal or small samples. The test begins by calculating the difference (d) between each observation and the hypothesized median. The absolute values of these differences are ranked in ascending order, with ties...
430
