Work

AIDeR: Precision-Oriented Author Name Matching for FAIR Scholarly Metadata Enrichment at Scale

Year: 2026

Type: article

Abstract: Matching personal names across heterogeneous bibliographic datasets is a fundamental yet challenging string matching problem: variations in formatting, cultural conventions, abbreviations, and transliteration make it difficult to reliably identify when two name strings refer to the same individual.This paper presents AIDeR (Author IDentifier Reconciler), a precision-oriented name matching algorithm that prioritizes high-confidence correspondences over broad recall, and that is deployed in production in the OpenAIRE Graph to enrich millions of scholarly records with verified ORCID identifiers.Rather than relying only on edit-distance or fuzzy similarity metrics, AIDeR employs a sequential pipeline of exact, token-based and fuzzy strategies, each targeting a well-defined class of name variation, and rejects matches that remain ambiguous after the pipeline has been drained. We describe the algorithm in detail and demonstrate its effectiveness in enriching Scholarly Knowledge Graphs with verified ORCID identifiers by aligning author name strings in publication metadata with names in ORCID profiles. We then build a large-scale benchmark from DBLP and the ORCID public data dump (more than 6 million candidate author–publication pairs) and compare AIDeR against widely used string matching baselines under three filtering scenarios that progressively isolate the contribution of the matcher from gaps in the ground truth. AIDeR consistently delivers the best F1 (up to 0.927) and reaches a precision of 0.998 in the strictest scenario, while no individual baseline matches both its precision and its recall. Both AIDeR and the full evaluation pipeline, driven by a Jupyter notebook that reproduces every figure and table inthis paper, are released as open source

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.23104071

Authors: Baglioni, Miriam, Bloisi, Giambattista, La Bruzzo, Sandro, Manghi, Paolo

Institutions: Istituto di Scienza e Tecnologie dell'Informazione "Alessandro Faedo", OpenAIRE Non-Profit Civil Partnership, National Research Council

Language: English

Topic: Data Quality and Management

Subfield: Management Science and Operations Research

Field: Decision Sciences

Domain: Social Sciences

Open Access status: green

Funder: European Commission

Awards: GraspOS: next Generation Research Assessment to Promote Open Science, EOSC Beyond: advancing innovation and collaboration for research