Privacy preserving record linkage

我们将对必须保护记录持有人隐私的情况下的记录链接算法进行研究。记录链接的目标是识别对应于同一实体的来自不同数据库的记录。随着当今数据库规模的不断扩大，链接过程的复杂性和缓慢的速度已经成为执行记录链接的主要挑战之一。此外，当包含机密信息的数据库跨不同组织链接时，保护数据的隐私和机密性问题变得至关重要。我们寻求设计和评估有效和准确地连接两个或多个数据库的方法，以确保不发生隐私泄露。效率很重要，因为数据库可能非常大。将来，加拿大公司可能会将该项目创建的Record Linkage算法应用于来自各种客户（包括健康、医疗和政府组织）的大量数据集合。来自这些组织的数据几乎总是包含机密信息。由于通常使用诸如记录持有人的姓名、地址和出生日期等通用识别数据来链接记录，因此保护个人隐私的方法是一个关键问题。我们还将尝试确定隐私水平与运营效率之间权衡的定量措施。

We will perform research on Record Linkage algorithms for cases where the privacy of record holders must be preserved. The goal of Record Linkage is to recognize records from different databases that correspond to the same entity. With the increasing size of today's databases, the complexity and slow speed of the linking process has become one of the major challenges for performing Record Linkage. Furthermore, when databases that contain confidential information are linked across different organizations, the issue of protecting the privacy and confidentiality of the data becomes crucial. We seek to devise and evaluate methods for efficient and accurate linking of two or more databases that ensure no privacy breach occurs. Efficiency is important because the databases may be very large. In the future, Canadian companies may apply the Record Linkage algorithms created by this project to large collections of data from a diversity of clients including health, medical, and governmental organizations. The data from such organizations almost always include confidential information. Since commonly identifying data, such as names, addresses, and date of births of record holders are typically used to link records, methods for preserving the privacy of individuals are a key concern. We will also try to identify quantitative measures of trade offs between levels of privacy versus efficiency of operations.