A fast 'parsing' algorithm for conceptual clustering
N.S. Jaishankar · 2002
The notion of conceptual clustering was introduced by Michalski in 1980 where he used concepts at a more physical level such as (color='red') and (height='tall') and so on. Shankar et al. explored the possibility of using a knowledge base in the clustering process. They used concepts at one or more levels higher than the physical level. The author suggests a method which uses only integers during the clustering process. The idea involves reorganizing the knowledge base (KB) used in Shankar et al. and the presentation of a method for clustering, which uses KB so re-organized. The crux of the reorganization involves associating every node in cohesion forest (CF) with a unique integer. This association is made in a 'breadth-first-traversal' fashion for all trees in CF.>