EU_parliament_sentiment_english

EU sentiment detection model

This model is trained to classify the sentiment of statements on the European Union. It finetuned a bert-base-german-cased model on 1700 sentences from the German parliament. It is trained to detect policy skepticism, hard EU skepticism, EU praisal, demand for more EU intergration and EU Speech.

Examples:

1. EU Speech: With today's motion, Die Linke has once again proven that it has no interest whatsoever in a serious European and foreign policy.

2. More EU: It is crucial that these things are regulated harmoniously at European level.

3. Policy Skepticism: Eurobarometer surveys show that certain institutions in Europe, especially bureaucratic organizations, are harshly criticized by the population.

4. EU praisal: Ladies and gentlemen, Europe is not just a peace project - it has always been a project to secure prosperity together and in solidarity.

5)Hard Euroscepticism: The EU juggernaut has become unreformable for a reason that you do not want to see.

Model Details Finetuned from model: google-bert/bert-based-cased Epochs: 5 F1 Weighted (5 categories): 0.5257

Lattmann, J. (2025, March 17). Detecting EU sentiment in texts: A LLM Machine Learning application for Euroscepticism research. https://doi.org/10.31219/osf.io/mravb_v1