Cliqueing‐a technique for producing maximally connected clusters
Gordon M. Gerson · Journal of the American Society for Information Science · 1978
Abstract A technique is demonstrated whereby a large data base may be automatically classified into maximally connected clusters. These are called cliques. The data base is initially clustered by MacQueen's “k‐means” algorithm. These clusters are then subjected to an enrichment process whereby through a transposition of the basic properties matrix, two‐directional connections may be shown. These are employed to the fullest extent with all connections being completed and references appearing in all cliques with which common or related references appear. The data base employed is a selection of US patents.