README.md

metadata

license: mit
language:
  - en
task_categories:
  - visual-document-retrieval
pretty_name: Author and Affiliation Parsing Benchmark Dataset
tags:
  - scholarly-documents
  - scientific-papers
  - author-extraction
  - affiliation-extraction
  - metadata-parsing
  - grobid
  - open-science
size_categories:
  - 10K<n<100K

Author and Affiliation Parsing Benchmark Dataset

This dataset provides a benchmark suite for evaluating author name parsing and author-affiliation linking systems in scientific documents.

The goal of this benchmark is to assess the ability of document parsing systems to accurately extract:

• author names
• author ordering
• affiliations

The benchmark combines several existing evaluation datasets covering different document types and scientific domains.

Dataset Overview

Scientific document parsing remains a challenging task due to the diversity of publication layouts, author representations, and affiliation formats.

The benchmark focuses specifically on the author and affiliation extraction task.

The benchmark is composed of the following subsets:

Subset	Source	Description
etd-author-affiliations	cometadata/etd-author-affiliations	Electronic theses and dissertations with annotated author and affiliation information
arxiv-author-affiliations	cometadata/arxiv-author-affiliations	arXiv scientific articles with author-affiliation annotations
PMC_sample_1943	GROBID evaluation dataset	Sample of PubMed Central articles used for GROBID evaluation
eLife_984	GROBID evaluation dataset	eLife articles benchmark subset
biorxiv-10k-test-2000	GROBID evaluation dataset	bioRxiv article benchmark subset
PLOS_1000	GROBID evaluation dataset	PLOS articles benchmark subset