1,309 labeled claims in Portuguese paired with professional fact-check classifications and source metadata. The data is sourced from fact-checking agencies via the ClaimReview schema to facilitate the study of misinformation in a non-English context.
Use Cases
- Develop classification models to identify misinformation in Portuguese using the claim text and label fields
- Perform linguistic analysis on fake news patterns within the 1,309 Portuguese language entries
- Benchmark automated fact-checking systems against professional human-labeled classifications
Strengths
- 1,309 unique claims with corresponding truth labels
- Sourced from professional fact-checking agencies using the ClaimReview structured data schema
- Includes news text, the associated fact-check, and a final classification for each entry