Knowledge Base Stratification and Merging Based on Degree of Support

  • Authors:

  • Anthony Hunter;Weiru Liu

  • Affiliations:

  • Department of Computer Science, University College London, London, UK WC1E 6BT;School of Electronics, Electrical Engineering and Computer Science, Queen's University Belfast, Belfast, UK BT7 1NN

  • Venue:
  • ECSQARU '09 Proceedings of the 10th European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty
  • Year:
  • 2009

  • Information fusion

    SUM'10 Proceedings of the 4th international conference on Scalable uncertainty management

Quantified Score

Hi-index 0.00

Visualization

Abstract

Most operators for merging multiple knowledge bases (where each is a set of formulae) aim to produce a knowledge base as output that best reflects the information available in the input. Whilst these operators have some valuable properties, they do not provide explicit information on the degree to which each formula in the output has been, in some sense, supported by the different knowledge bases in the input. To address this, in this paper, we first define the degree of support that a formula receives from input knowledge bases. We then provide two ways of determining formulae which have the highest degree of support in the current collection of formulae in KBs, each of which gives a preference (or priority) over formulae that can be used to stratify the formulae in the output. We formulate these two preference criteria, and present an algorithm that given a set of knowledge bases as input, generates a stratified knowledge base as output. Following this, we define some merging operators based on the stratified base. Logical properties of these operators are investigated and a criterion for selecting merging operators is introduced.