A review of author name disambiguation techniques for the PubMed bibliographic database

Abstract

Author names in bibliographic databases often suffer from ambiguity owing to the same author appearing under different names and multiple authors possessing similar names. It creates difficulty in associating a scholarly work with the person who wrote it, thereby introducing inaccuracy in credit attribution, bibliometric analysis, search-by-author in a digital library and expert discovery. A plethora of techniques for disambiguation of author names has been proposed in the literature. In this article, we focus on the research efforts targeted to disambiguate author names specifically in the PubMed bibliographic database. We believe this concentrated review will be useful to the research community because it discusses techniques applied to a very large real database that is actively used worldwide. We make a comprehensive survey of the existing author name disambiguation (AND) approaches that have been applied to the PubMed database: we organise the approaches into a taxonomy; describe the major characteristics of each approach including its performance, strengths, and limitations; and perform a comparative analysis of them. We also identify the datasets from PubMed that are publicly available for researchers to evaluate AND algorithms. Finally, we outline a few directions for future work.

Model Reuse in Machine Learning for Author Name Disambiguation: An Exploration of Transfer Learning

“…Biomedical researcher names have also received much attention from disambiguation scholars. The public MEDLINE database has been one of the most frequent target datasets for AND [6]. Maintained by the National Library of Medicine (NLM), MEDLINE contains the largest collection of publication records in biomedicine.…”

“…Maintained by the National Library of Medicine (NLM), MEDLINE contains the largest collection of publication records in biomedicine. 6 We matched author names in MEDLINE (December 2015 version) to ORCID ids following the steps outlined in [21]. That process resulted in approximately 2M ORCID ids-linked instances in papers published between 1991~2014.…”

“…KISTI data includes 41,605 author name instances representing 6,937 computer scientists. The data are semimanually labeled by researchers at the Korea Institute of 6 https://www.nlm.nih.gov/bsd/medline.html 7 https://journals.aps.org/datasets Science Technology and Information (KISTI) and Kyungsung University [35].…”

“…Among many methods for author name disambiguation (AND), machine-learning-based approaches have been reported to produce better disambiguation results than rulebased algorithms or heuristics [3][4][5]. This might be because machine learning generates models adapted to each of diverse name ambiguity settings and scenarios by detecting sophisticated disambiguation patterns through training and validation on labeled data [6]. Although crucial in machine learning for AND, creating labeled data is time-consuming and costly.…”

“…When 'author name disambiguation' is searched on Google Scholar, about 60% of the 200 papers that have received at least one citation rely on machine learning techniques. AND machine learning approaches can be divided into two general groups -classification studies and clustering studies -based on the types of learning output they produce [3,6,9]. The first is focused on producing disambiguation results at a pair level.…”