NER-Modell 20 des Projekts Dehmel Digital

This dataset contains two types of resources: Firstly, three Named Entity Recognition models developed in the context of the project "Dehmel digital" for the automatic annotation of persons, places, artworks and organisations in letters from the period around 1900. The training corpus for model 20 consists of 400,000 manually annotated tokens.
Second, a table in which the results of the performance test are broken down in detail. The performance was calculated on the basis of eight different test texts, each consisting of 10,000 manually annotated tokens.