Scientific knowledge role transition prediction from a knowledge hierarchical structure perspective

Jinqing Yang and Jiming Hu

Journal of Informetrics, 2025, vol. 19, issue 1

Abstract: There are several potential patterns in the evolution of scientific knowledge. In order to delve deeper into the changes in function and role during the evolution of knowledge, we have proposed a research framework that examines the transition of scientific knowledge roles from the perspective of a hierarchical structure. We constructed two classification models of transition possibility and transition type to predict whether one undergoes a role transition and which type of role transition it belongs to. Several datasets were constructed by utilizing the entire corpus of publications available in PubMed and the history records of MeSH. Among the tasks of transition type prediction and transition possibility prediction, the Gradient Boosting classifier performed the best. The binary classification model of transition possibility achieved a precision of 72.58 %, a recall of 71.04 %, and an F1 score of 71.78 %. The multi-classification model of transition possibility had a macro-F1 score of 61.29 %, a micro-F1 score of 84.07 %, and a weighted-F1 score of 82.90 %. Further, we found that the knowledge genealogy features contribute the most to the prediction of transition possibility while knowledge attribute and network structure features have a significantly greater influence on the prediction of transition type. Most features have an obvious effect on the role transition of the Content-change type, followed by Child-generation and Localization-shift types.

Keywords: Knowledge evolution; Knowledge role transition; Machine learning; Knowledge network; Knowledge tree (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S175115772400124X
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:infome:v:19:y:2025:i:1:s175115772400124x

DOI: 10.1016/j.joi.2024.101612

Access Statistics for this article

Journal of Informetrics is currently edited by Leo Egghe

More articles in Journal of Informetrics from Elsevier
Bibliographic data for series maintained by Catherine Liu ().