Information inconsistencies detection using a rule-map technique

  • Authors:
  • Jun Ma;Jie Lu;Guangquan Zhang

  • Affiliations:
  • Faculty of Engineering Information Technology, University of Technology, Sydney (UTS), P.O. Box 123, Broadway, NSW 2007, Australia;Faculty of Engineering Information Technology, University of Technology, Sydney (UTS), P.O. Box 123, Broadway, NSW 2007, Australia;Faculty of Engineering Information Technology, University of Technology, Sydney (UTS), P.O. Box 123, Broadway, NSW 2007, Australia

  • Venue:
  • Expert Systems with Applications: An International Journal
  • Year:
  • 2009

Quantified Score

Hi-index 12.05

Visualization

Abstract

Timely detecting information inconsistencies (anomalies) in real-time information provides strong support for decision-making in a dynamic decision-making situation. Existing techniques for information inconsistencies detection mainly focus on stored information by using a single structured-fixed descriptive model which always requires support from sufficient prior knowledge. The aim of this study is to develop a method for information inconsistencies detection for real-time information in dynamic decision-making situation where prior knowledge is insufficient by using multiple descriptive models. First, a rule-map technique is presented. A rule-map is a hierarchical directed graph, whose vertexes are selected descriptive models and whose arcs represent the covering relationship between descriptive models. A rule-map provides a strategy for selecting detecting descriptive models by means of the covering relationship and its structure is adjustable with the change in a situation. Then, a real-time information inconsistencies detection method, named RMDID, is developed based on the rule-map technique, which can take full advantage of multiple descriptive models. Finally, the proposed RMDID method is tested through two real cases. Experiments indicate that the proposed rule-map technique can trace the changes of a dynamic decision-making situation and the developed RMDID method can efficiently detect potential anomalies in real-time information.