| 1 |
German Climate Change Tweet Corpus (GerCCT) ...
|
|
|
|
Abstract:
First release of the GerCCT Corpus, a German tweet resource annotated for argument components, argument properties, sarcasm and toxic language. The corpus consists of 1,200 tweets and its annotations. Each tweet is associated with its respective source tweet, i.e. the tweet it replies to. Source tweets were used to provide annotators with additional context. The annotations refer to the reply tweet, i.e. NOT to the source tweet. For copyright reasons we cannot distribute the actual tweet content. Instead we share the source and reply tweet IDs and the annotations. The current version includes class annotations on the document level, i.e. on the tweet level. We are working on creating the respective span annotations. ...
|
|
Keyword:
argument mining; computational linguistics; German; Twitter
|
|
URL: https://dx.doi.org/10.5281/zenodo.6479491
https://zenodo.org/record/6479491
|
|
BASE
|
|
Hide details
|
|
|
|
|
