Combining multiple scholarly relationships with author cocitation analysis: A preliminary exploration on improving knowledge domain mappings

Abstract

Abstract Author cocitation analysis (ACA) is a branch of bibliometrics and knowledge representation that aims to map knowledge domains. However, ACA has been criticized because count-based measurement is too simple, and resulting maps are insufficiently informative. Since different scholarly relationships, e.g., coauthorship and author bibliographic coupling relationships, can extract out different relationships among authors in various perspectives, combining them with ACA for constructing knowledge domain mappings is our major purpose. The proposed method constructs the hybrid matrix from all relationships in four steps: relationship normalization, calculating the similarity between scholarly relationships, calculating adjustment parameters, and constructing hybrid relationships. The important parameters for integrating these matrices are calculated according to the distance in the hyperspace transformed from the similarity among the scholarly relationships by exploratory factor analysis. Compared with ACA, the results of the proposed method show: (1) More sub-fields in the given discipline can be identified when combining other scholarly relationships; (2) The more scholarly relationships added into ACA, the more details in terms of research area the method will find; (3) Good visualization in clustering is depicted when we combine other scholarly relationships. As a result, the proposed method offers a good choice to understand researchers and to map knowledge domains in a study field for integrating more scholarly relationships at the same time.