Euronews: Europeana Newspapers Collection

Designed for named entity recognition tasks, containing text in multiple languages including Dutch, French, and German. The annotations were generated by experts, and the data is stored in Parquet format. Specific details on row count, column structure, and data instances are unavailable.

Use Cases

• Perform token classification for named entity recognition on multilingual text data.
• Analyze the distribution of named entities across different languages such as Dutch, French, and German.
• Train a multilingual NER model using expert-generated annotations as supervision.

Strengths

• Annotations are expert-generated, which can indicate high label quality.
• The dataset is multilingual, supporting languages including Dutch, French, and German.
• Data is provided in Parquet format, which is efficient for storage and processing.

Limitations

• Critical structural information like row count, column names, and data splits is unknown.
• The dataset size is categorized as 'n1 K', suggesting it may contain only a few thousand instances.
• The source description and curation rationale are marked as 'More Information Needed', limiting understanding of data scope.

Provenance

Source

community-datasets on Hugging Face, originating from Europeana Newspapers.

Freshness

The dataset was last updated on 2024-06-24.

Geography

Region is tagged as 'us', but the dataset relates to European newspapers and multilingual content.