[Air-L] Datasets for stance detection

[kiran gvr]

Hi,

Stance detection has been well studied in the context of debate sites of
various kinds (both monologic and dialogic) and for student essays.

Here's a few of the resources that come to my mind, that have ground truth
labels. Let me know if you want more.

1. The TU Darmstadt argumentation datasets -- look for the Argument
Annotated Essays (there are two versions)
https://www.ukp.tu-darmstadt.de/data/argumentation-mining/

2. The internet argument corpus contains huge dumps of debate data
https://nlds.soe.ucsc.edu/iac2 (there are again two versions)

3. The group at UT Dallas has done some work on stance classification and
reasoning. Their dataset can be found here:
http://www.hlt.utdallas.edu/~saidul/stance/stance.html

Regards,
Kiran

On Wed, Jul 12, 2017 at 6:05 AM, Ophélie Fraisier <ophelie.fraisier at irit.fr>
wrote:

> Hello,
>
> I’m currently working on stance detection on social media, and I’m looking
> for datasets to evaluate my model.
> I have several datasets of tweets concerning political events, but I would
> like to also consider other platforms.
> Ideally, I would need datasets with textual content, users profiles, their
> ground truth stances and a non negligible amount of interactions (of any
> form) between the users, but I would take any recommandation :)
>
> Thanks for your suggestions.
>
> ———
> Ophélie Fraisier
> PhD Student in opinion mining on social media
> CEA Tech Toulouse - IRIT, IRIS team
>
> _______________________________________________
> The Air-L at listserv.aoir.org mailing list
> is provided by the Association of Internet Researchers http://aoir.org
> Subscribe, change options or unsubscribe at: http://listserv.aoir.org/
> listinfo.cgi/air-l-aoir.org
>
> Join the Association of Internet Researchers:
> http://www.aoir.org/




-- 

--
Kiran