Czech Historical Named Entity Corpus v 1.0

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 4458–4465 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 4458 Czech Historical Named Entity Corpus v 1.0 Helena Hubkov´a♣, Pavel Kr´al♣, Eva Pettersson♠ ♣Department of Computer Science and Engineering, Faculty of Applied Sciences, University of West Bohemia, Univerzitn´ı 2732/8, 301 00 Pilsen, Czech Republic ♠Dept. of Linguistics and Philology, Uppsala University P.O. Box 256, SE-751 05 Uppsala, Sweden ♣{hhubkova, pkral}@kiv.zcu.cz, ♠eva.pettersson@lingfil.uu.se Abstract As the number of digitized archival documents increases very rapidly, named entity recognition (NER) in historical documents has become very important for information extraction and data mining. For this task an annotated corpus is needed, which has up to now been missing for Czech. In this paper we present a new annotated data collection for historical NER, composed of Czech historical newspapers. This corpus is freely available for research purposes at http://chnec.kiv.zcu.cz/. For this corpus, we have defined relevant domain-specific named entity types and created an annotation manual for corpus labelling. We further conducted some experiments on this corpus using recurrent neural networks in order to in order to show baseline results on this dataset. We experimented with randomly initialized embeddings and static and dynamic fastText word embeddings. We achieved 0.73 F1 score with a bidirectional LSTM model using static fastText embeddings. Keywords: Historical Czech, Historical Named Entity Corpus, LSTM, Named Entity Recognition, Neural Networks 1. Introduction Named entity recognition (NER) is a fundamental task in natural language processing (NLP). As the amount of digitized archival material has increased rapidly dur- ing the last few decades, NER has become an im- portant step for information extraction in the field of historical document analysis. However, a lack of annotated historical data for NER is an obstacle to research in this area. Therefore this paper intro- duces a new annotated data collection dedicated to historical NER with some experiments using meth- ods based on neural networks. The experiments are conducted to show baseline results on this dataset. The corpus is freely available for research purposes at http://chnec.kiv.zcu.cz/. We also plan to submit this corpus to be a part of the Language Research In- frastructure of LINDAT/CLARIN project. This research is carried out within the framework of the project Modern Access to Historical Sources, presented in the Porta Fontium portal.1 One goal of this project is to enable intelligent full-text access to the printed historical documents from the Czech-Bavarian border region. Accordingly, our original data sources to create the corpus are scanned texts from a Czech historical newspaper Posel od ˇCerchova from the second half of the 19th century. The scanned materials were digitized by optical char- acter recognition (OCR). Then, we defined specific named entities based on the project purpose itself in combination with named-entity types from ˇSevˇc´ıkov´a et al. (2007b). We also created an annotation manual for the final Czech historical named entity corpus. For our experiments, we created a bidirectional LSTM model for sequence-labelling inspired by Chiu and 1http://www.portafontium.eu/ Nichols …

Content truncated to 500 words