Historical Species Mentions in 19th-Century Württemberg Texts

Accuracy results for four time spans evaluate an automated LLM-based method for extracting species mentions from historical texts. The method achieved 92.6% recall and 95.3% precision for species detection, with 83.0% accuracy for linking mentions to GBIF identifiers. This 5.5 KB dataset, created by Maximilian C. Teich, demonstrates a scalable approach for generating biodiversity datasets from archival material.

Use Cases

• Benchmark the 92.6% recall and 95.3% precision metrics for species detection in historical text across four time spans.
• Analyze the 83.0% accuracy rate for linking extracted species mentions to GBIF database identifiers.
• Compare the performance of the automated LLM-based extraction method for different time spans within the 19th-century Württemberg context.
• Use the accuracy results to validate and refine automated pipelines for generating biodiversity datasets from historical sources.

Strengths

• Provides specific evaluation metrics: 92.6% recall, 95.3% precision for detection, and 83.0% accuracy for species linking.
• Focuses on a defined historical and geographical context: 19th-century Württemberg, Germany.
• Dataset is openly licensed under CC BY 4.0, facilitating reuse and sharing.

Limitations

• The dataset is very small at 5.5 KB, containing only accuracy results rather than the underlying extracted species data.
• The scope is limited to evaluating a specific method on texts from a single region and century, limiting generalizability.
• Row and column counts are unknown, making it difficult to assess the granularity of the accuracy results.

Provenance

Source

figshare

Collection Method

Accuracy results derived from an automated approach using prompted large language models to detect and link species mentions in historical texts.

Time Range

19th century

Freshness

Last updated on 2026-03-24.

Geography

Württemberg, Germany