The AffectEventKB data set contains 571,424 unique events. Each event is also associated with a polarity label (highest score) and a distribution over three polarity classes (positive, negative, and neutral), which were automatically assigned using a weakly supervised learning algorithm.
The EventPolarity data set contains events with manually annotated affective polarity labels. The HumanNeeds data set contains affective events with positive or negative polarity that have additional manual annotations for human need category labels, which indicate the type of reason for the affective polarity.
These resources can be obtained using this link: http://www.cs.utah.edu/~hbding/affectEventResource.html [www.cs.utah.edu]
We hope that these resources may be useful for deeper narrative story understanding and implicit sentiment analysis. If you have any questions, please contact us: Haibo Ding, hbding at cs.utah.edu Ellen Riloff, riloff at cs.utah.edu -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/html Size: 2384 bytes Desc: not available URL: <https://mailman.uib.no/public/corpora/attachments/20180619/5e522380/attachment.txt>