On a real-world cancer registry linkage task, fine-tuned Mistral-7B made only 6 classification errors among 52,917 record pairs, while embedding-based blocking reduced candidate pairs by 92% with a small recall loss.
Record linkage: Current practice and future directions
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Leveraging Language Models for Automated Patient Record Linkage
On a real-world cancer registry linkage task, fine-tuned Mistral-7B made only 6 classification errors among 52,917 record pairs, while embedding-based blocking reduced candidate pairs by 92% with a small recall loss.