Linking named entities in Dutch historical newspapers

Abstract

We improved access to the collection of Dutch historical newspapers of the Koninklijke Bibliotheek by linking named entities in the newspaper articles to corresponding Wikidata descriptions by means of machine learning techniques and crowdsourcing. Indexing the Wikidata identifiers for named entities together with the newspaper articles opens up new possibilities for retrieving articles that mention these resources and searching the newspaper collection using semantic relations from Wikidata. In this paper we describe our steps so far in setting up this combination of entity linking, machine learning and crowdsourcing in our research environment as well as our planned activities aimed at improving the quality of the links and extending the semantic search capabilities.