Incremental structuring of knowledge bases

Abstract

An important structuring mechanism for knowledge bases is building an inheritance hierarchy of classes based on the content of their knowledge objects. The hierarchy can be used to handle several query processing tasks more efficiently. Building and maintaining this hierarchy is a difficult task for the knowledge engineer. The notion of knowledge space has been previously proposed to help automate such a task. In this paper an incremental algorithm for building the knowledge space is proposed and tested on a sample knowledge base. The empirical behavior shows that after a point the knowledge space can be updated in close to constant time on the average. The knowledge space structure is presented in the framework of the theory of concept lattices. The presentation suggests that the notion of a complete knowledge lattice might be a worthwhile alternative because of the additional richness of the structure. However, this richness induces a large amount of additional processing and storage...