LLM-powered ParlaCAP dataset analyzes European parliamentary agenda setting

Researchers have developed ParlaCAP, a new dataset designed to analyze parliamentary agenda setting across 28 European countries. This dataset utilizes a multilingual LLM in a teacher-student framework to create domain-specific policy topic classifiers, achieving annotation agreement comparable to human annotators. The ParlaCAP dataset includes extensive metadata and sentiment predictions, enabling comparative research on political attention and representation. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Enables new avenues for comparative political science research using LLM-based analysis of parliamentary data.

RANK_REASON This is a research paper introducing a new dataset and methodology for analyzing parliamentary discourse.

Read on arXiv cs.CL →

• ParlaCAP
• European Parliaments
• LLM
• Comparative Agendas Project
• ParlaMint
• ParlaSent
• arXiv

• paper
• other