A Dataset of Historical Medical Periodicals Annotated with Textual Genre

Abstract

Historical corpora, especially those compiled from magazines and periodicals, are complex due to the diversity of text types and evolving genre conventions. Addressing these challenges requires systematic genre annotation and well-defined classification schemes to support downstream NLP tasks. This paper introduces a dataset of historical medical periodical texts in German and Swedish annotated for textual genre and additional features that may influence genre identification, such as the presence of OCR errors. We describe the development of the genre classification, annotator recruitment and training procedures, and provide an analysis of the annotator agreement.

Resources

Details