A method to assist in the proper management of rule bases in Web information systems

Abstract

Purpose One key issue of maintaining Web information systems is to guarantee the consistency of their knowledge base, in particular, the rules governing them. There are currently few methods that can ensure that rule bases management can scale to the amount of knowledge in these systems environment. Design/methodology/approach In this paper, the authors propose a method to detect correct dependencies between rules. This work represents a preliminary step for a proposal to eliminate rule base anomalies. The authors previously developed a method that aimed to ameliorate the extraction of rules dependency relationships using a new technique. In this paper, they extend the proposal with other techniques to increase the number of extracted rules dependency relationships. The authors also add some modules to filter and represent them. Findings The authors evaluated their own method against other semantic methods. The results show that this work succeeded in extracting better numbers of correct rules dependency relationships. They also noticed that the rule groups deduced from this method’s results are very close to those provided by the rule bases developers. Originality/value This work can be applied to knowledge bases that include a fact base and a rule base. In addition, it is independent of the field of application.