Sign in to view source links and access this dataset
Description
12,800 short statements labeled for truthfulness, sourced from PolitiFact.com's API. Each statement was evaluated by a PolitiFact editor, and the label distribution is relatively balanced, with instances for most labels ranging from 2,063 to 2,638 cases. The dataset was uploaded by 'ucsbai' and last updated on Hugging Face in January 2024.
Use Cases
Train text classifiers for fake news detection based on labeled short statements.
Analyze patterns in political misinformation based on editor-provided truthfulness judgments.
Develop models to ground truthfulness assessments based on the accompanying analysis reports.
Strengths
12,800 human-labeled statements provide a substantial corpus for model training.
Label distribution is relatively balanced, with most categories containing between 2,063 and 2,638 instances.
Each judgment includes a lengthy analysis report from an editor, providing context for the label.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Last updated 2024-01-18 11:08:08; freshness should be verified.
Provenance
Source
PolitiFact.com's API
Collection Method
Statements were evaluated and labeled by PolitiFact.com editors.
Freshness
Last updated 2024-01-18 11:08:08.
License is unknown; terms of use must be verified before application.