Powered by OpenAIRE graph
Found an issue? Give us feedback
addClaim

Named entity disambiguation in biomedical literature

Authors: Gittings, Matthew;

Named entity disambiguation in biomedical literature

Abstract

La présence de conflits d'intérêts dans la littérature biomédicale est un champ d'intérêt croissant en santé publique, en raison des préoccupations en lien avec les entités qui génèrent des résultats de recherche à leur avantage. Néanmoins, les conflits d'intérêts et l'influence potentielle qu'ils exercent sur les résultats de recherche demeurent un phénomène difficile à saisir et à étudier à grande échelle. Dans cette thèse, nous évaluerons un ensemble de données de chercheurs et leurs engagements avec des entités, extraits de la divulgation d'un conflit d'intérêts d'article, et des méthodes actuelles pour déchiffrer les références textuelles à des entités du monde réel. Nous présentons d'abord un modèle de classification pour lier les mentions d'entités localement dans le même article, puis nous passerons à la démystification des mentions d'entités à travers les articles. Nous ajusterons ensuite et appliquerons ces méthodes à l'ensemble de notre base de données, ce qui donnera un vaste réseau d'entités, qui peut être exploité dans l'étude des conflits d'intérêts. Enfin, nous présenterons cette analyse où nous caractérisons l'engagement de l'industrie à travers la lentille de l'influence des chercheurs. Nous énoncerons que la liaison d'entités locales peut être effectuée avec une grande précision, et que la désambiguïsation d'entité globale, bien que plus impliquée et nécessitant plus de réglage, peut également être réalisée à un haut niveau. Nous présentons aussi des preuves provenant d'une analyse de réseaux qui montre que les entreprises ciblent les chercheurs ayant des niveaux d'influence plus élevés dans leurs domaines que les autres types d'organisations gouvernementales et de fondations.

The presence of conflict of interest in biomedical literature is a space of increasing interest in public health, due to concerns of entities driving research outcomes for their benefit. However, conflict of interest, and the potential influence that it exerts on research outcomes, remains a difficult phenomenon to capture and investigate on a large scale. In this thesis, we present and characterize a dataset of researchers and their engagements with entities, extracted from article conflict of interest disclosures, and present methods to disambiguate the textual references to real-world entities We first present a classification model for linking entity mentions locally in the same article, and then move on to disambiguating entity mentions globally from across articles. We then tune and apply these methods to our entire dataset, yielding a large researcher-entity network, that can be leveraged in the study of conflict of interest. Finally, we present one such analysis where we characterize industry engagement through the lens of researcher influence We find that the task of local entity linking can be performed with high accuracy, and that global entity disambiguation, though more involved and requiring more tuning, can be performed to a high degree as well. We also present evidence from a network analysis that shows that corporations target researchers with higher levels of influence in their fields than other entity types of government organizations and foundations

Ruths, Derek (Internal/Supervisor)

Country
Canada
Related Organizations
Keywords

Computer Science

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
0
Average
Average
Average
Related to Research communities
Upload OA version
Are you the author of this publication? Upload your Open Access version to Zenodo!
It’s fast and easy, just two clicks!