Data mining model based on multi-agent for the intelligent distributed framework. (1st October 2010)

• Record Type: Journal Article
• Authors: Mateo, Romeo Mark A.; Lee, Jaewan
• Is Part Of: International journal of intelligent information and database systems. Volume 4:Number 4(2010)
• Page Start: 322
• Page End: 336
• Publication Date: 2010-10-01
• Subjects: data mining -- multi-agent systems -- MAS -- distributed systems -- agent-based systems -- intelligent modelling -- clustering -- object grouping -- classification -- association mining -- simulation -- delay time -- load distribution; Database management -- Computer programs -- Periodicals; Information retrieval -- Computer programs -- Periodicals; Information storage and retrieval systems -- Computer programs -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Intelligent agents (Computer software) -- Periodicals; 006.33
• Languages: English
• ISSNs: 1751-5858

Abstract

Most researches in large-scale distributed system only focus on improving a single scheme, and the relationships between the schemes that affect the performance of each scheme are ignored. In this paper, an intelligent distributed framework is introduced to address the use of intelligent models for the adaptive schemes of a distributed system. This paper tackles two aspects; the general model of an agent-based data mining to implement the schemes of distributed object system using data mining algorithms and efficient interactions of the schemes using multi-agent approach. The adaptive schemes use clustering to construct classes for object grouping, classification method to classify the requests using the classes constructed from clustering, and association mining to generate rules for predicting the next object needed to be replicated. These schemes are provided with relationships for the adaptive technique and the interaction is based on the action model of multi-agent system. Simulation result shows significant improvements on serving clients by minimised delay time and efficient load distribution.