Developing a scientific knowledge graph through conceptual linking of academic classifications

Abstract

This paper describes the process of the construction of a scientific knowledge graph via semantic annotation and linking of academic research fields. The method aims to combine the broad scope of generic classifications of research areas with the specialized knowledge and domain fields of specific scientific classifications in order to build a unified scientific knowledge graph including all the research fields of the respective scientific areas in a common hierarchy. First, a survey of scientific classifications has been conducted in order to identify the ones that are the most common and complete for the knowledge graph to build upon. Classifications from different research domains have been used to semantically annotate their thematic topics and fields. The various scientific fields are connected based on their similarity, enlightening and creating, in this way, cross-domain research fields. A core scientific graph containing main fields of science with their relations to domain specific vocabularies and classifications has been created. This knowledge graph can be used to retrieve specific scientific fields in a related, broader or narrower, research area. Finally, a use case leveraging the semantic features of the graph is presented, indicating its usefulness for research activities and PhD web services.