Examining Communities in the Transdisciplinary Area of Cognitive Science: Automatic Classification for Examining Communities in the Web of Science Using Unsupervised Clustering Methods

Maxime Sainte-Marie, Laura Ridenour, Vincent Lariviére

Abstract


We propose methodology for examining classification to identify and make explicit community perspectives that are neglected by traditional journal-subject classification in order to provide a more flexible and customizable classification system. Our method is based on keyword matches, and is applied to the broad transdisciplinary area of cognitive science. In the Web of Science (WoS), Scopus, and the National Science Foundation (NSF) classification, the classification of journals places each journal into a silo based on pre-determined categories deemed appropriate to demonstrate the relatedness of journals. Classification at the journal level does not necessarily represent the perspectives of a community, as a community in both membership and topical scope may transcend the bounds of a single journal classification. Our approach is novel because we examine topics within the transdisciplinary domain of cognitive science, and within that domain, we identify community perspectives on the conceptual contents as found in the titles of publications in the WoS.

Keywords


Automatic classification, community-based classification, unsupervised clustering methods, knowledge organization

Full Text:

PDF


DOI: http://dx.doi.org/10.7152/acro.v29i1.15463