Cluster-lift method for mapping research activities over a concept tree

Abstract

The paper builds on the idea by R. Michalski of inferential concept interpretation for knowledge transmutation within a knowledge structure taken here to be a concept tree. We present a method for representing research activities within a research organization by doubly generalizing them. To be specific, we concentrate on the Computer Sciences area represented by the ACM Computing Classification System (ACM-CCS). Our cluster-lift method involves two generalization steps: one on the level of individual activities (clustering) and the other on the concept structure level (lifting). Clusters are extracted from the data on similarity between ACM-CCS topics according to the working in the organization. Lifting leads to conceptual generalization of the clusters in terms of "head subjects" on the upper levels of ACM-CCS accompanied by their gaps and offshoots. A real-world example of the representation is provided. © 2010 Springer-Verlag Berlin Heidelberg.