LIARArg: Fake News Detection Data Enhanced with Argumentative Features

A dataset for fake news detection, enhanced by argumentative features, shared by author Xiaoou Wang. The dataset is available as a 46.8 MB CSV file under a CC-BY-4.0 license and was last updated on May 12, 2026. Its canonical archived version is hosted on the Nakala platform.

Use Cases

• Training classifiers for fake news detection based on argumentative features.
• Analyzing the relationship between argument structure and content veracity.
• Benchmarking natural language processing models on a task combining fact-checking and argument mining.

Strengths

• Dataset is 46.8 MB in size, indicating a substantial text corpus.
• Explicitly includes argumentative features, a specific enhancement for the detection task.
• Published under a permissive CC-BY-4.0 license, facilitating reuse and sharing.

Limitations

• Row count is unknown, which may limit suitability assessment for large-scale training.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

figshare; canonical version on Nakala (doi:10.34847/NKL.2CEC5407).

Freshness

Last updated 2026-05-12 11:43:22; freshness should be verified.