An approach to discover and recommend cross-domain bridge-keywords in document banks

Abstract

Purpose - The co-word analysis method is commonly used to cluster-related keywords into the same keyword domain. In other words, traditional co-word analysis cannot cluster the same keywords into more than one keyword domain, and disregards the multi-domain property of keywords. The purpose of this paper is to propose an innovative keyword co-citation approach called "Complete Keyword Pair (CKP) method", which groups complete keyword sets of reference papers into clusters, and thus finds keywords belonging to more than one keyword domain, namely bridge-keywords. Design/methodology/approach - The approach regards complete author keywords of a paper as a complete keyword set to compute the relations among keywords. Any two complete keyword sets whose corresponding papers are co-referenced by the same paper are recorded as a CKP. A clustering method is performed with the correlation matrix computed from the frequency counts of the CKPs, for clustering the complete keyword sets. Since keywords may be involved in more than one complete keyword set, the same keywords may end up appearing in different clusters. Findings - Results of this study show that the CKP method can discover bridge-keywords with average precision of 80 per cent in the Journal of the Association for Computing Machinery citation bank during 2000-2006 when compared against the benchmark of Association for Computing Machinery Computing Classification System. Originality/value - Traditional co-word analysis focuses on co-occurrence of keywords, and therefore, cannot cluster the same keywords into more than one keyword domain. The CKP approach considers complete author keyword sets of reference papers to discover bridge-keywords. Therefore, the keyword recommendation system based on CKP can recommend keywords across multiple keyword domains via the bridge-keywords.