Fine-tuned BERT predicts Austrian MPs' parties from speeches
Original title: Analysing Austrian Parliamentary Speeches by Predicting Party Affiliations
Governing parties grow more alike in speech, opposition diverges, and switchers adapt only after changing party.
Researchers fine-tune BERT to assign each Austrian National Council speech a probability distribution over parties, then aggregate scores by speaker and party. They treat mis-assignments as signals of political similarity rather than model error. Coalition partners become more alike, opposition parties move away from governing parties, and members who switch parties usually change speech style only afterward. Party-affiliation predictions for speeches from 1996 to 2018 are released publicly.
Unconfirmed: Findings are the authors' reported results and are not independently verified here.
- The study covers Austrian National Council speeches from 1996 to 2018.
- Each speech receives a probability distribution over Austrian parties.
- Mis-predicted parties are used to study similarity and political change.
- Members who switch party usually adapt speech only after the switch.
- The authors publicly release the prediction dataset.
- Organizations
- Austrian National Council
- Countries
- Austria
- AI models
- BERT
From the original article
1 min read at arxiv.orgThis paper examines speech characteristics of the Austrian National Council members (CMs). We introduce analysis methodology based on a fine-tuned BERT-Model which assigns to every speech a probability distribution over Austrian parties. These are then aggregated at the speaker and party level.
By Sarah Sulollari · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.