This paper has been withdrawn by Noah Dasanaike
[Submitted on 29 Jan 2026 (v1), last revised 15 Oct 2026 (this version, v3)]
No PDF available, click to view other formats
Abstract:Record linkage, the process of matching records that refer to the same entity across datasets, is essential to empirical social science but remains methodologically underdeveloped. Researchers treat it as a preprocessing step, applying ad hoc rules without quantifying the uncertainty that linkage errors introduce into downstream analyses. Existing methods either achieve low accuracy or require substantial labeled training data. I present EnsembleLink, a method that achieves high accuracy without any training labels. EnsembleLink leverages pre-trained language models that have learned semantic relationships (e.g., that “South Ozone Park” is a neighborhood in “New York City” or that “Lutte ouvriere” refers to the Trotskyist “Workers’ Struggle” party) from large text corpora. On benchmarks spanning city names, person names, organizations, multilingual political parties, and bibliographic records, EnsembleLink matches or exceeds methods requiring extensive labeling. The method runs locally on open-source models, requiring no external API calls, and completes typical linkage tasks in minutes.
Submission history
From: Noah Dasanaike [view email]
[v1] Thu, 29 Jan 2026 00:41:11 UTC (14 KB)
[v2] Fri, 27 Feb 2026 22:46:29 UTC (15 KB)
[v3] Thu, 15 Oct 2026 01:58:41 UTC (1 KB) (withdrawn)
