A labeled collection of short statements for fake news detection. The dataset is sourced from PolitiFact and contains statements categorized for veracity. It is hosted on Kaggle and tagged under 'News'.
Use Cases
- Classify statement_veracity using text features from the statement content.
- Analyze patterns in false statements versus true statements using labeled categories.
- Train models like BERT or LSTMs for binary or multi-class text classification on political claims.
- Perform feature engineering on statement text to identify linguistic markers of deception.
Strengths
- Dataset focuses on a specific, high-impact domain of political fact-checking.
- Provides labeled data for a supervised learning task in NLP.
Limitations
- Unknown row count limits assessment of model training potential.
- Unknown column details prevent evaluation of feature richness or auxiliary metadata.
- Temporal coverage and update recency are unspecified, which may affect model relevance.
Provenance
- Source
- PolitiFact
- Collection Method
- null
- Time Range
- null
- Freshness
- null
- Geography
- null