Dataset De Argumentenfabriek in English

This dataset contains the argument graphs of <em>De Argumentenfabriek</em> [1], originally created as PDF files in Dutch. The dataset refers to the paper "<em>Fine and Coarse Granular Argument Classification before Clustering</em>" and contains 365 graphs with more than 11,000 arguments, which were manually converted into a CSV and translated into English using DeepL [2] in order to be able to reconstruct their stances, frames, and clusterings.<br> Moreover, there is also a concentrated dataset consisting of 49 graphs comprising 2,186 arguments, which was used in the evaluation of the paper. Two further files contain the unified stance and frame labels of the arguments.<br> <br> [1]: https://www.argumentenfabriek.nl/<br> [2]: https://www.deepl.com/