Understanding the impact of three derived text formats on authorship classification with Delta

Due to copyright law, Text and Data Mining with copyrighted texts faces a lot of restrictions in terms of storage, publication and follow-up use of the resulting corpora, which, however, is against the spirit of open data in digital humanities. As a solution to the problem, the concept of derived text formats (DTFs) have been suggested and discussed. The presented paper did an empirical study by transforming texts into token-based DTFs and provide a review on the usefulness of the transformed texts on authorship classification. The results show that selectively reducing information on individual tokens could ensure, to a certain extent, that the authorship classification results are not affected too much. Ein Beitrag zur 9. Tagung des Verbands "Digital Humanities im deutschsprachigen Raum" - DHd 2023 Open Humanities Open Culture.