Impresso Mediaagencies Ner Dataset: Annotated Historical News Agency Mentions

v0.1 data is derived from legacy French/German HIPE-style news-agency annotations, converted to JSONL, and manually reviewed. The dataset provides curated token-classification data for news-agency and radio-station mentions in Impresso historical newspaper text. It was created by the impresso-project and last updated on June 2, 2026.

Use Cases

• Train named entity recognition models based on annotated canonical media-source organization mentions.
• Evaluate model performance on historical text based on manually reviewed dev/test disagreements.
• Study the evolution of media agency reporting based on mentions in historical newspaper text.

Strengths

• Data underwent manual review against model disagreements and was updated according to annotation guidelines v2.0.
• Annotations are derived from a legacy, established HIPE-style annotation framework for French and German text.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

impresso-project

Collection Method

Derived from legacy HIPE-style annotations, converted to JSONL, and manually reviewed.

Freshness

Last updated 2026-06-02 06:34:52; freshness should be verified.