A Benchmark of Named Entity Recognition Approaches in Historical Documents Application to 19th Century French Directories - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2022

A Benchmark of Named Entity Recognition Approaches in Historical Documents Application to 19th Century French Directories

Résumé

Named entity recognition (NER) is a necessary step in many pipelines targeting historical documents. Indeed, such natural language processing techniques identify which class each text token belongs to, e.g. "person name", "location", "number". Introducing a new public dataset built from 19th century French directories, we first assess how noisy modern, off-the-shelf OCR are. Then, we compare modern CNNand Transformer-based NER techniques which can be reasonably used in the context of historical document analysis. We measure their requirements in terms of training data, the effects of OCR noise on their performance, and show how Transformer-based NER can benefit from unsupervised pre-training and supervised fine-tuning on noisy data. Results can be reproduced using resources available at https://github.com/ soduco/paper-ner-bench-das22.
Fichier principal
Vignette du fichier
preprint_das_2022_abadie_carlinet_chazalon_dumenieu.pdf (949.02 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03698609 , version 1 (18-06-2022)

Identifiants

Citer

N. Abadie, E. Carlinet, J. Chazalon, B. Duménieu. A Benchmark of Named Entity Recognition Approaches in Historical Documents Application to 19th Century French Directories. Document Analysis Systems. DAS 2022., May 2022, La Rochelle, France. ⟨10.1007/978-3-031-06555-2_30⟩. ⟨hal-03698609⟩

Relations

94 Consultations
320 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More