 |
    |
An Agent-Based Data Mining System for Ontology Evolution
Maja Hadzic and Darshan Dillon
Digital Ecosystems and Business Intelligence Institute, Curtin University of Technology GPO Box U1987, Perth 6845, Australiam.hadzic@curtin.edu.au
darshan.dillon@curtin.edu.au
Abstract. We have developed an evidence-based mental health ontological model that represents mental health in multiple dimensions. The ongoing addition of new mental health knowledge requires a continual update of the Mental Health Ontology. In this paper, we describe how the ontology evolution can be realized using a multi-agent system in combination with data mining algorithms. We use the TICSA methodology to design this multi-agent system which is composed of four different types of agents: Information agent, Data Warehouse agent, Data Mining agents and Ontology agent. We use UML 2.1 sequence diagrams to model the collaborative nature of the agents and a UML 2.1 composite structure diagram to model the structure of individual agents. The Mental Heath Ontology has the potential to underpin various mental health research experiments of a collaborative nature which are greatly needed in times of increasing mental distress and illness.
Keywords: ontology evolution, data mining, multi-agent system, multi-agent system design, mental health, mental health ontology
LNCS 5872, p. 836 ff.
lncs@springer.com
© Springer-Verlag Berlin Heidelberg 2009