writer stance dataset

Repository: writer stance - 1000 German texts from the X-stance corpus (see References X-stance) annotated for writer stance<br> - two formats: conll and BIO (sort of) - see below<br> - goal: explicit and implicit writer stance wrt. to entities (nouns) and (some) events (verbs)<br> - implicit and explicit are not distinguished (no indication which case was annotated) - labels: BIO Conll<br> - in favour: PRO p<br> - against: CON c<br> - neutral: O no label<br> - 3 Annotations (see References DeInStance) Formats: conll: (parzu parser) 1 Das die ART ART Def|Neut|Nom|Sg 2 det _ _<br> p2 Arbeitsgesetz Arbeitsgesetz N NN Neut|Nom|Sg 3 subj _ _<br> 3 regelt regeln V VVFIN 3|Sg|Pres|Ind 0 root _ _<br> 4 die die ART ART Def|Fem|Acc|Pl 5 det _ _<br> 5 Arbeitszeiten Arbeitszeit N NN Fem|Acc|Pl 3 obja _ _<br> 6 und und KON KON _ 3 kon _ _<br> 7 schützt schützen V VVFIN _|_|Pres|Ind 6 cj _ _<br> 8 den die ART ART Def|Masc|Acc|Sg 9 det _ _<br> p9 Arbeitnehmer Arbeitnehmer N NN Masc|Acc|Sg 7 obja _ _<br> 10 . . $. $. _ 0 root _ _ head was annotated directly in front of the index (1. column)<br> e.g. "p2 Arbeitsgesetz Arbeitsgesetz N NN Neut|Nom|Sg 3 subj _ _"<br> i.e. the writer is in favour of "Arbeitsgesetz" BIO: wordform lemma label Das die O<br> Arbeitsgesetz Arbeitsgesetz PRO<br> regelt regeln O<br> die die O<br> Arbeitszeiten Arbeitszeit O<br> und und O<br> schützt schützen O<br> den die O<br> Arbeitnehmer Arbeitnehmer PRO<br> . . O head was annotated (last column)<br> e.g. "Arbeitsgesetz Arbeitsgesetz PRO" <br> for X-stance corpus see also https://vamvas.ch/more-general-stance-detection-with-x-stance <br> References: @article{X-stance,<br> author = {Jannis Vamvas and<br> Rico Sennrich},<br> title = {X-Stance: {A} Multilingual Multi-Target Dataset for Stance Detection},<br> journal = {CoRR},<br> volume = {abs/2003.08385},<br> year = {2020},<br> url = {https://arxiv.org/abs/2003.08385},<br> eprinttype = {arXiv},<br> eprint = {2003.08385},<br> timestamp = {Tue, 24 Mar 2020 16:42:29 +0100},<br> biburl = {https://dblp.org/rec/journals/corr/abs-2003-08385.bib},<br> bibsource = {dblp computer science bibliography, https://dblp.org}<br> } @inproceedings{DeInStance,<br> booktitle = {17th Conference on Natural Language Processing (KONVENS)},<br> month = {September},<br> title = {DeInStance: Creating and Evaluating a {G}erman Corpus for Fine-Grained Inferred Stance Detection},<br> author = {Anne Gohring and Manfred Klenner and Sophia Conrad},<br> publisher = {ACL Anthology},<br> year = {2021},<br> pages = {213--217},<br> language = {english},<br> url = {https://doi.org/10.5167/uzh-207940},<br> abstract = {We introduce deInStance, a corpus of 1000 politicians? answers in German (de) containing sentences labeled with explicitly expressed and inferred stances - pro and con relations - by 3 annotators. They achieved an acceptable inter-rater agreement given the inherent subjective nature of the task. A first baseline, a fine-tuned BERT-based token classifier, achieved F1-scores of around 70\% . Our focus is on the difficult subclass of sentences comprising only non-polar words, but still with an (implicit) pro or con perspective of the writer.}<br> }