SNF project sentiment inference (finished Dec 2022)

• German sentiment inference based on connotation frames 
• identify writer, reader and text perspective 
• factuality determination for inference  validation
• role framing: who gets positively/negatively conceptualized
• attitude prediction: who is in favour of /against who

• Anne Göhring 
• Alison Yong-Ju Kim
• Manfred Klenner
• Dylan Massey

Former members:

• Noëmi Aepli
• Sophia Conrad
• Andreas Säuberli 

The project was funded by the Swiss National Science Fondation ( 105215-179302) and  was finished  December 2022.

publications

• Animacy Classification for German (TGZ, 57 KB)
• Animacy Classification Gold Standard Data (TGZ, 116 KB)
• Pro/Con Sentiment Inference Gold Standard Data (TGZ, 227 KB)
• Quantifed NPs Silver Standard Data (TGZ, 267 KB)
• Negative Entities Quantified: Gold Standard Data (TGZ, 85 KB)
• X-stance Writer Perspective: Gold Standard Data (TGZ, 1 MB)

Also available from Zenodo

A demo of the current implementation is available here