Sentiment Inference: Pro and Contra relation dataset

500 German sentences annotated for pro/con relations and polar roles of entities (negative/positive actors/effects): see References for a conceptual introduction. files: annotator1.conll .. annotator3.conll format: conll (parzu parser) with annotations - annotations at the end of the conll parse tree<br> - c = con<br> - p = pro<br> - neff,peff = negative, positive effect<br> - nac, pac = negative, positive actor<br> - the head indices are used for annotation (see below)<br> - c1,6 = Hofstetter con Gewerkschaften<br> - neff6 = negative Effekt on Gewerkschaften <br> Note: in these annotations, pro/con is not an intentional relation - in "Snow blocks the driveway" it holds: con(snow,driveway)<br> - "snow" is a negative element wrt. to driveway<br> - use our animacy classifier to identify those case with an actor (see References lrec, available via IGGSA download) <br> Example:<br> 1 Hofstetter Hofstetter N NE _|Nom|Sg 2 subj _ _ <br> 2 wirft werfen V VVFIN 3|Sg|Pres|Ind 0 root _ _ <br> 3 im in PREP APPRART Dat 2 pp _ _ <br> 4 Interview Interview N NN Neut|Dat|Sg 3 pn _ _ <br> 5 den die ART ART Def|Fem|Dat|Pl 6 det _ _ <br> 6 Gewerkschaften Gewerkschaft N NN Fem|Dat|Pl 2 objd _ _ <br> 7 vor vor PTKVZ PTKVZ _ 2 avz _ _ <br> 8 , , $, $, _ 0 root _ _ <br> 9 sie sie PRO PPER 3|Pl|_|Nom 10 subj _ _ <br> 10 wollen wollen V VMFIN 3|Pl|Pres|_ 2 s _ _ <br> 11 die die ART ART Def|Fem|_|Sg 12 det _ _ <br> 12 Branche Branche N NN Fem|_|Sg 13 obja _ _ <br> 13 anschwärzen anschwärzen V VVINF _ 10 aux _ _ <br> 14 . . $. $. _ 0 root _ _ <br> c1,6<br> p1,12<br> neff6 <br> References: @inproceedings{stance,<br> booktitle = {LSDSem 2017/LSD-Sem Linking Models of Lexical, Sentential and Discourse-level Semantics},<br> month = {April},<br> title = {Stance Detection in Facebook Posts of a German Right-wing Party},<br> author = {Manfred Klenner and Don Tuggener and Simon Clematide},<br> publisher = {ResearchBib},<br> year = {2017},<br> language = {english},<br> url = {https://doi.org/10.5167/uzh-136567}<br> }<br> @inproceedings{perspectives,<br> booktitle = {18th International Conference on Computational Linguistics and Intelligent Text Processing},<br> month = {April},<br> title = {Verb-mediated Composition of Attitude Relations Comprising Reader and Writer Perspective},<br> author = {Manfred Klenner and Simon Clematide and Don Tuggener},<br> publisher = {ResearchBib},<br> year = {2017},<br> language = {english},<br> url = {https://doi.org/10.5167/uzh-136569},<br> doi = {10.1007/978-3-319-77116-8\_11}<br> }<br> @inproceedings{harmonization,<br> booktitle = {Proceedings of the 5th Swiss Text Analytics Conference (SwissText) \&amp; 16th Conference on Natural Language Processing (KONVENS)},<br> editor = {Sarah Ebling and Don Tuggener and Manuela H{\"u}rlimann and Martin Volk},<br> month = {Juni 2020},<br> title = {Harmonization Sometimes Harms},<br> author = {Manfred Klenner and Anne G{\"o}hring and Michael Amsler},<br> publisher = {Virtual Event}<br> year = {2020},<br> language = {english},<br> url = {https://doi.org/10.5167/uzh-197961}<br> }<br> @inproceedings{lrec,<br> month = {Juni},<br> author = {Manfred Klenner and Anne G{\"o}hring},<br> booktitle = {Proceedings of the Language Resources and Evaluation Conference},<br> address = {Marseille, France},<br> title = {Animacy Denoting {G}erman Nouns: Annotation and Classification},<br> publisher = {European Language Resources Association},<br> pages = {1360--1364},<br> year = {2022},<br> language = {english},<br> url = {https://doi.org/10.5167/uzh-219148},<br> abstract = {In this paper, we introduce a gold standard for animacy detection comprising almost 14,500 German nouns that might be used to denote either animate entities or non-animate entities. We present inter-annotator agreement of our crowd-sourced seed annotations (9,000 nouns) and discuss the results of machine learning models applied to this data.}<br> }<br>