ESNERA: Empirical and semantic named entity alignment for named entity dataset merging

TL;DR

This paper introduces an automatic label alignment method combining empirical and semantic similarities to merge NER datasets effectively, improving performance especially in low-resource domains.

Contribution

It proposes a novel, scalable, and interpretable label alignment approach for merging NER datasets using a greedy pairwise strategy based on label similarity.

Findings

Successfully merged multiple NER datasets with minimal performance loss.

Enhanced NER performance in the financial domain with limited data.

Demonstrated scalability and interpretability of the proposed method.

Abstract

Named Entity Recognition (NER) is a fundamental task in natural language processing. It remains a research hotspot due to its wide applicability across domains. Although recent advances in deep learning have significantly improved NER performance, they rely heavily on large, high-quality annotated datasets. However, building these datasets is expensive and time-consuming, posing a major bottleneck for further research. Current dataset merging approaches mainly focus on strategies like manual label mapping or constructing label graphs, which lack interpretability and scalability. To address this, we propose an automatic label alignment method based on label similarity. The method combines empirical and semantic similarities, using a greedy pairwise merging strategy to unify label spaces across different datasets. Experiments are conducted in two stages: first, merging three existing NER datasets into a unified corpus with minimal impact on NER performance; second, integrating this corpus with a small-scale, self-built dataset in the financial domain. The results show that our method enables effective dataset merging and enhances NER performance in the low-resource financial domain. This study presents an efficient, interpretable, and scalable solution for integrating multi-source NER corpora.