Replication Data for: Training and Evaluation of Named Entity Recognition Models for Classical Latin

Repository for NER data for Latin. The repo includes manual Named Entity Recognition (NER) annotation for three texts: Cicero, In M. Antonium Oratio Philippica prima (3754 tokens, annotated by Margherita Fantoli) Juvenal, Satirae 1-3 (4399 tokens, annotated by Evelien De Graaf and Margherita Fantoli) Tacitus, Historiae 1 (11983 tokens; annotated by Evelien De Graaf) The tokens are taken from the LASLA corpus. Every token is associated to its token and lemma URIs as found in the LiLa Knowledge Base. Latin_Silver_Data: This folder contains the automatic NER annotation for the texts included in the portion of the LASLA corpus linked to the LiLa Knowledge Base, and includes the URIs of the linking. The model used for the annotation is found at retrain with best params. See the following paper about the LiLa-LASLA linking: Linking the LASLA Corpus in the LiLa Knowledge Base of Interoperable Linguistic Resources for Latin (Fantoli et al., LDL 2022) Latin NER code: contains the code to train a LatinBERT model for Latin NER