Batch re-linking #410
Description
This issue was auto-created as part of the Zenhub sub-issues migration. The original entity was opened on 2025-02-12. For more details see this post.
Summary
Define and implement a process for re-linking the MPI, or portions of it, on a periodic basis.
Objectives
- Discovery and research helping define Record Linker's options for re-linking
- A proposed strategy for how Record Linker can solve this problem
- Implementation of any new code needed to support the new process
- Documentation for our users on how to perform this process
Background / Context
Batch re-linkage can help mitigate linkage drift, which may be challenging for small person clusters but becomes resolvable as data volume increases. However, as the size of an MPI grows, periodically re-evaluating the entire cluster becomes increasingly complex. To address this, it’s crucial to strike a balance between the frequency of re-linking and the volume of data being re-evaluated.