This respository is a supplement to a presentation held at the 2022 TEI Conference in Newcastle on "How to Represent Topic Models in Digital Scholarly Editions" by Ulrike Henny-Krahmer and Frederike Neuber. Slides and abstract of the talk are available on Zenodo. The repository provides data and code that illustrate how topic models could be integrated into digital scholarly editions by transforming the results of a modeling with Mallet to XML/TEI.
- tei-examples: example data from the digital edition Correspondence of ‘Jean Paul’s surroundings’
- mallet-output: results of applying topic modeling to the Jean Paul corpus by using Mallet.
- scripts: code for transforming Mallet results to XML/TEI
- topics-with-words-to-taxonomy.xsl: creates a tei:taxonomy with topics and their most probable words
- topics-in-texts-to-keywords.xsl: records the probablity of the occurence topics in a document in tei:keywords
- topic-state-to-text.xsl: embedds references to topics on word level in the document text by using tei:w [this is work-in-progress]
- text-to-html.xsl: transforms XML with embedded topic references to HTML
- html-examples: examplary HTML output, in which topic words are highlighted directly in the edition text
The content of this repository is available under CC BY-SA 4.0 License.
For reusing any content of the repository, please cite us as follows:
Henny-Krahmer, Ulrike and Frederike Neuber. 2022. “Topic Models in TEI.” GitHub.com. https://github.com/hennyu/topic-model-tei.