Sample data for evaluating Scholix relationship SubTypes for linked data publications

Scholix links provide a standardized framework for establishing connections between research publications and their associated datasets or related data publications, thereby fostering improved discoverability, reusability, and reproducibility of research data.This dataset aims to facilitate the evaluation of the degree of relatedness between literature publications and their associated linked data publications. It comprises 3,600 tuples, each representing a pair of a literature publication (A) and a linked data publication (B) connected through Scholix links.

Dataset Contents

1. Scholix Links: The dataset includes 450 Scholix links for each of the eight most frequently observed relationship types between literature and linked data publications, as expressed in the "RelationshipType - SubType" field of Scholix metadata:

IsSupplementedBy

IsReferencedBy

IsRelatedTo

References

Documents

Cites

IsSupplementTo

IsCitedBy

2. Publication Metadata: In addition to the Scholix links, the dataset is augmented with metadata for each publication, including titles and author names. This metadata was harvested from the Crossref and DataCite APIs.

3. Relatedness Measures: To estimate the degree of relatedness between literature and linked data publications, the dataset includes numeric measures for the similarity of authors' lists and publication titles for each tuple.

Data Sources

Scholix links were harvested from the Scholexplorer API.

Publication metadata (titles and author names) were obtained from the Crossref and DataCite APIs.

This dataset can be valuable for researchers and practitioners working on linked literature and data publications, evaluating the quality of existing links, or developing algorithms to identify related publications across different domains.