Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

From raw affiliations to organization identifiers

Myrto Kallipoliti, Serafeim Chatzopoulos, Miriam Baglioni, Eleni Adamidi, Paris Koloveas, Thanasis Vergoulis

TL;DR

This work tackles the challenge of linking free-text affiliation strings to organization identifiers by introducing AffRo, a rule-based, keyword-driven matching pipeline, and AffRoDB, a thoroughly expert-curated benchmark. AffRo processes affiliations through preprocessing, matching, and disambiguation phases, employing targeted pruning, partitioning, and a two-stage similarity-based scoring to handle multi-organization strings and noisy data. The authors provide extensive experimental evaluation against baselines (S2AFF, OpenAlex) on both the expert-curated AffRoDB and Crossref-derived data, showing superior precision, recall, and F1 in most settings, and they deliver an open API and open-source implementation for broad reuse. Their publicly available dataset, update policy, and integration into OpenAIRE Graph underscore practical impact for improving metadata quality, enabling richer bibliometric analyses, and enhancing interoperability across scholarly data ecosystems.

Abstract

Accurate affiliation matching, which links affiliation strings to standardized organization identifiers, is critical for improving research metadata quality, facilitating comprehensive bibliometric analyses, and supporting data interoperability across scholarly knowledge bases. Existing approaches fail to handle the complexity of affiliation strings that often include mentions of multiple organizations or extraneous information. In this paper, we present AffRo, a novel approach designed to address these challenges, leveraging advanced parsing and disambiguation techniques. We also introduce AffRoDB, an expert-curated dataset to systematically evaluate affiliation matching algorithms, ensuring robust benchmarking. Results demonstrate the effectiveness of AffRp in accurately identifying organizations from complex affiliation strings.

From raw affiliations to organization identifiers

TL;DR

This work tackles the challenge of linking free-text affiliation strings to organization identifiers by introducing AffRo, a rule-based, keyword-driven matching pipeline, and AffRoDB, a thoroughly expert-curated benchmark. AffRo processes affiliations through preprocessing, matching, and disambiguation phases, employing targeted pruning, partitioning, and a two-stage similarity-based scoring to handle multi-organization strings and noisy data. The authors provide extensive experimental evaluation against baselines (S2AFF, OpenAlex) on both the expert-curated AffRoDB and Crossref-derived data, showing superior precision, recall, and F1 in most settings, and they deliver an open API and open-source implementation for broad reuse. Their publicly available dataset, update policy, and integration into OpenAIRE Graph underscore practical impact for improving metadata quality, enabling richer bibliometric analyses, and enhancing interoperability across scholarly data ecosystems.

Abstract

Accurate affiliation matching, which links affiliation strings to standardized organization identifiers, is critical for improving research metadata quality, facilitating comprehensive bibliometric analyses, and supporting data interoperability across scholarly knowledge bases. Existing approaches fail to handle the complexity of affiliation strings that often include mentions of multiple organizations or extraneous information. In this paper, we present AffRo, a novel approach designed to address these challenges, leveraging advanced parsing and disambiguation techniques. We also introduce AffRoDB, an expert-curated dataset to systematically evaluate affiliation matching algorithms, ensuring robust benchmarking. Results demonstrate the effectiveness of AffRp in accurately identifying organizations from complex affiliation strings.
Paper Structure (31 sections, 3 equations, 4 figures, 3 tables)

This paper contains 31 sections, 3 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Background & Problem definition
  3. Methodology
  4. High-level overview
  5. Preprocessing phase
  6. Cleaning and stemming.
  7. Keyword labeling, partitioning, and exact matching.
  8. Partition pruning.
  9. String shortening for universities and entity names.
  10. Matching phase
  11. Candidate identification.
  12. Results refinement.
  13. Results disambiguation
  14. Expert-curated dataset
  15. Dataset creation process
  16. ...and 16 more sections

Figures (4)

  • Figure 1: Example of AffRo's key processing steps.
  • Figure 2: Performance metrics varying the $\texttt{window}$ parameter.
  • Figure 3: Performance metrics varying similarity thresholds $\texttt{sim\_o}$ and $\texttt{sim\_u}$.
  • Figure :

Theorems & Definitions (3)

  • definition 1: PID
  • definition 2: Affiliation matching
  • definition 3: Cosine similarity