Simple Demographics Often Identify People Uniquely

作者: Latanya Sweeney

DOI: 10.1184/R1/6625769.V1

关键词:

摘要: In this document, I report on experiments conducted using 1990 U.S. Census summary data to determine how many individuals within geographically situated populations had combinations of demographic values that occurred infrequently. It was found few characteristics often combine in uniquely or nearly identify some individuals. Clearly, released containing such information about these should not be considered anonymous. Yet, health and other person-specific are publicly available form. Here surprising results only three fields information, even though typical releases contain more fields. 87% (216 million 248 million) the population United States reported likely made them unique based {5-digit ZIP, gender, date birth}. About half (132 53%) identified by {place, birth}, where place is basically city, town, municipality which person resides. And at county level, {county, birth} 18% population. general, needed a person.

参考文章(4)
Latanya Sweeney, Weaving Technology and Policy Together to Maintain Confidentiality Journal of Law Medicine & Ethics. ,vol. 25, pp. 98- 110 ,(1997) , 10.1111/J.1748-720X.1997.TB01885.X
G.W. Smith, Modeling security-relevant data semantics ieee symposium on security and privacy. pp. 384- 391 ,(1990) , 10.1109/RISP.1990.63866
Kenneth P. Bogart, Introductory combinatorics ,(1983)